Tag Archives: Personal Finance – Career & Education

What the evolution of our own brains can tell us about the future of AI

Posted on by

The explosive growth in artificial intelligence in recent years — crowned with the meteoric rise of generative AI chatbots like ChatGPT — has seen the technology take on many tasks that, formerly, only human minds could handle. But despite their increasingly capable linguistic computations, these machine learning systems remain surprisingly inept at making the sorts of cognitive leaps and logical deductions that even the average teenager can consistently get right. 

In this week's Hitting the Books excerpt, A Brief History of Intelligence: Evolution, AI, and the Five Breakthroughs That Made Our Brains, AI entrepreneur Max Bennett explores the quizzical gap in computer competency by exploring the development of the organic machine AIs are modeled after: the human brain. 

Focusing on the five evolutionary "breakthroughs," amidst myriad genetic dead ends and unsuccessful offshoots, that led our species to our modern minds, Bennett also shows that the same advancements that took humanity eons to evolve can be adapted to help guide development of the AI technologies of tomorrow. In the excerpt below, we take a look at how generative AI systems like GPT-3 are built to mimic the predictive functions of the neocortex, but still can't quite get a grasp on the vagaries of human speech.

It's a picture of a brain with words over it
HarperCollins

Excerpted from A Brief History of Intelligence: Evolution, AI, and the Five Breakthroughs That Made Our Brains by Max Bennett. Published by Mariner Books. Copyright © 2023 by Max Bennett. All rights reserved.

Words Without Inner Worlds

GPT-3 is given word after word, sentence after sentence, paragraph after paragraph. During this long training process, it tries to predict the next word in any of these long streams of words. And with each prediction, the weights of its gargantuan neural network are nudged ever so slightly toward the right answer. Do this an astronomical number of times, and eventually GPT-3 can automatically predict the next word based on a prior sentence or paragraph. In principle, this captures at least some fundamental aspect of how language works in the human brain. Consider how automatic it is for you to predict the next symbol in the following phrases:

  • One plus one equals _____

  • Roses are red, violets are _____

You’ve seen similar sentences endless times, so your neocortical machinery automatically predicts what word comes next. What makes GPT-3 impressive, however, is not that it just predicts the next word of a sequence it has seen a million times — that could be accomplished with nothing more than memorizing sentences. What is impressive is that GPT-3 can be given a novel sequence that it has never seen before and still accurately predict the next word. This, too, clearly captures something that the human brain can _____.

Could you predict that the next word was do? I’m guessing you could, even though you had never seen that exact sentence before. The point is that both GPT-3 and the neocortical areas for language seem to be engaging in prediction. Both can generalize past experiences, apply them to new sentences, and guess what comes next.

GPT-3 and similar language models demonstrate how a web of neurons can reasonably capture the rules of grammar, syntax, and context if it is given sufficient time to learn. But while this shows that prediction is part of the mechanisms of language, does this mean that prediction is all there is to human language? Try to finish these four questions:

  • If 3x + 1 = 3, then x equals _____

  • I am in my windowless basement, and I look toward the sky, and I see _____

  • He threw the baseball 100 feet above my head, I reached my hand up to catch it, jumped, and _____

  • I am driving as fast as I can to LA from New York. One hour after passing through Chicago, I finally _____

Here something different happens. In the first question, you likely paused and performed some mental arithmetic before being able to answer the question. In the other questions, you probably, even for only a split second, paused to visualize yourself in a basement looking upward, and realized what you would see is the ceiling. Or you visualized yourself trying to catch a baseball a hundred feet above your head. Or you imagined yourself one hour past Chicago and tried to find where you would be on a mental map of America. With these types of questions, more is happening in your brain than merely the automatic prediction of words.

We have, of course, already explored this phenomenon—it is simulating. In these questions, you are rendering an inner simulation, either of shifting values in a series of algebraic operations or of a three-dimensional basement. And the answers to the questions are to be found only in the rules and structure of your inner simulated world.

I gave the same four questions to GPT-3; here are its responses (responses of GPT-3 are bolded and underlined):

  • If 3x + 1 = 3 , then x equals 1

  • I am in my windowless basement, and I look toward the sky, and I see a light, and I know that it is a star, and I am happy.

  • He threw the baseball 100 feet above my head, I reached my hand up to catch it, jumped, and caught it. It was a lot of fun!

  • I am driving as fast as I can to LA from New York. One hour after passing through Chicago, I finally get to see the Pacific Ocean.

All four of these responses demonstrate that GPT-3, as of June 2022, lacked an understanding of even simple aspects of how the world works. If 3x + 1 = 3, then x equals 2/3, not 1. If you were in a basement and looked toward the sky, you would see your ceiling, not stars. If you tried to catch a ball 100 feet above your head, you would not catch the ball. If you were driving to LA from New York and you’d passed through Chicago one hour ago, you would not yet be at the coast. GPT-3’s answers lacked common sense.

What I found was not surprising or novel; it is well known that modern AI systems, including these new supercharged language models, struggle with such questions. But that’s the point: Even a model trained on the entire corpus of the internet, running up millions of dollars in server costs — requiring acres of computers on some unknown server farm — still struggles to answer common sense questions, those presumably answerable by even a middle-school human.

Of course, reasoning about things by simulating also comes with problems. Suppose I asked you the following question:

Tom W. is meek and keeps to himself. He likes soft music and wears glasses. Which profession is Tom W. more likely to be?

1) Librarian

2) Construction worker

If you are like most people, you answered librarian. But this is wrong. Humans tend to ignore base rates—did you consider the base number of construction workers compared to librarians? There are probably one hundred times more construction workers than librarians. And because of this, even if 95 percent of librarians are meek and only 5 percent of construction workers are meek, there still will be far more meek construction workers than meek librarians. Thus, if Tom is meek, he is still more likely to be a construction worker than a librarian.

The idea that the neocortex works by rendering an inner simulation and that this is how humans tend to reason about things explains why humans consistently get questions like this wrong. We imagine a meek person and compare that to an imagined librarian and an imagined construction worker. Who does the meek person seem more like? The librarian. Behavioral economists call this the representative heuristic. This is the origin of many forms of unconscious bias. If you heard a story of someone robbing your friend, you can’t help but render an imagined scene of the robbery, and you can’t help but fill in the robbers. What do the robbers look like to you? What are they wearing? What race are they? How old are they? This is a downside of reasoning by simulating — we fill in characters and scenes, often missing the true causal and statistical relationships between things.

It is with questions that require simulation where language in the human brain diverges from language in GPT-3. Math is a great example of this. The foundation of math begins with declarative labeling. You hold up two fingers or two stones or two sticks, engage in shared attention with a student, and label it two. You do the same thing with three of each and label it three. Just as with verbs (e.g., running and sleeping), in math we label operations (e.g., add and subtract). We can thereby construct sentences representing mathematical operations: three add one.

Humans don’t learn math the way GPT-3 learns math. Indeed, humans don’t learn language the way GPT-3 learns language. Children do not simply listen to endless sequences of words until they can predict what comes next. They are shown an object, engage in a hardwired nonverbal mechanism of shared attention, and then the object is given a name. The foundation of language learning is not sequence learning but the tethering of symbols to components of a child’s already present inner simulation.

A human brain, but not GPT-3, can check the answers to mathematical operations using mental simulation. If you add one to three using your fingers, you notice that you always get the thing that was previously labeled four.

You don’t even need to check such things on your actual fingers; you can imagine these operations. This ability to find the answers to things by simulating relies on the fact that our inner simulation is an accurate rendering of reality. When I mentally imagine adding one finger to three fingers, then count the fingers in my head, I count four. There is no reason why that must be the case in my imaginary world. But it is. Similarly, when I ask you what you see when you look toward the ceiling in your basement, you answer correctly because the three-dimensional house you constructed in your head obeys the laws of physics (you can’t see through the ceiling), and hence it is obvious to you that the ceiling of the basement is necessarily between you and the sky. The neocortex evolved long before words, already wired to render a simulated world that captures an incredibly vast and accurate set of physical rules and attributes of the actual world.

To be fair, GPT-3 can, in fact, answer many math questions correctly. GPT-3 will be able to answer 1 + 1 =___ because it has seen that sequence a billion times. When you answer the same question without thinking, you are answering it the way GPT-3 would. But when you think about why 1 + 1 =, when you prove it to yourself again by mentally imagining the operation of adding one thing to another thing and getting back two things, then you know that 1 + 1 = 2 in a way that GPT-3 does not.

The human brain contains both a language prediction system and an inner simulation. The best evidence for the idea that we have both these systems are experiments pitting one system against the other. Consider the cognitive reflection test, designed to evaluate someone’s ability to inhibit her reflexive response (e.g., habitual word predictions) and instead actively think about the answer (e.g., invoke an inner simulation to reason about it):

Question 1: A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?

If you are like most people, your instinct, without thinking about it, is to answer ten cents. But if you thought about this question, you would realize this is wrong; the answer is five cents. Similarly:

Question 2: If it takes 5 machines 5 minutes to make 5 widgets, how long would it take 100 machines to make 100 widgets?

Here again, if you are like most people, your instinct is to say “One hundred minutes,” but if you think about it, you would realize the answer is still five minutes.

And indeed, as of December 2022, GPT-3 got both of these questions wrong in exactly the same way people do, GPT-3 answered ten cents to the first question, and one hundred minutes to the second question.

The point is that human brains have an automatic system for predicting words (one probably similar, at least in principle, to models like GPT-3) and an inner simulation. Much of what makes human language powerful is not the syntax of it, but its ability to give us the necessary information to render a simulation about it and, crucially, to use these sequences of words to render the same inner simulation as other humans around us.

This article originally appeared on Engadget at https://www.engadget.com/hitting-the-books-a-brief-history-of-intelligence-max-bennett-mariner-books-143058118.html?src=rss

Assassin’s Creed Mirage review: A warm, bloody hug from an old friend

Posted on by

Editor's note: This article contains mild spoilers for Assassin's Creed Mirage.

The deeper I got into Assassin’s Creed Mirage, the more a sense of warm nostalgia washed over me. It felt like a cozy hug from an old friend. A comforting, bloody embrace.

The latest entry in Ubisoft's long-running open-world adventure franchise takes the series back to its roots. Mirage mostly forgoes the RPG approach Ubisoft adopted in the last three main games: Assassin's Creed Origins, Odyssey and Valhalla. I'd only played the latter of those and it didn't click for me, largely because of Ubisoft's propensity to ovestuff its games and partially because it strayed so far away from the earlier titles.

Some of Valhalla's DNA carries over to Mirage, which shouldn't be surprising as the latest game was originally envisioned as an expansion to the last 100-plus-hour epic. There is some loot to hunt for in the form of swords, daggers and outfits that give protagonist Basim some small upgrades, such as reducing the level of notoriety he gains while carrying out illegal actions or passively regenerating some health. These items are upgradable, as are your tools. One neat, if unrealistic perk, makes an enemy disintegrate after Basim eliminates them with a throwing knife. So, you can tweak your build to fit your playstyle to a certain degree.

A hooded figure prepares to kill an enemy with a blade protruding from a bracer in Assassin's Creed Mirage.
Ubisoft

There are skill trees too, but rather than unlocking things like a slight increase to the damage Basim deals, the abilities here are genuinely impactful. Pinpointing opponents and important items from further away, reducing fall damage and a chain assassination ability are all super useful tools for Basim to have in his belt.

Ubisoft has pulled back quite a bit on the RPG elements of the previous few games. You won’t be using bows, shields or two-handed weapons as you might in Valhalla, for instance. Still, there's just enough customization for folks who want to optimize (or min/max) Basim for the way they like to play.

"Just enough" is a thought I kept coming back to in the 17 hours it took me to beat the main story. Mirage is just the right length. There are just enough collectibles and side-quests to make the world feel rich but not overwhelming. There's just enough to the story, which is fairly by-the-numbers though gets more intriguing in the last couple of hours. There's just enough variety to the enemies.

There are only a few enemy types, and I love that Mirage doesn't go down the well-worn and nonsensical path of arbitrarily making them stronger based on their geographical location — an aspect of Dead Island 2 I greatly disliked. Although Basim largely has to make do with his sword and dagger (and, of course, the Hidden Blade), enemies have a variety of weapons. A trio of goons will pose a different threat when they have spears instead of swords. You'll have to navigate that melange of weaponry carefully, especially so when enemies surround you. Putting an onus on that and the level design for encounters helps make Mirage feel like more of a refreshing throwback.

A hooded figure prepares to drive his sword through an enemy in Assassin's Creed Mirage.
Ubisoft

In the main missions, I only encountered one traditional boss fight toward the end of the story. Practically every other enemy was susceptible to a single-button slaying. I absolutely made the most of that by sneaking up on assassination targets or distracting them with noise-making devices. The game actually discourages open combat, anyway. You won't gain experience points by killing tons of enemies. Staying stealthy is usually the way to go — unless you're a completionist, since there's a trophy/achievement that requires you to stay in open combat for 10 minutes. Thankfully, the game makes it fairly easy for you to slink around.

Contrary to my first impressions, the guards of Baghdad aren't all that smart. They'll often be briefly puzzled when they encounter the dead body of a colleague they were chatting with seconds earlier before walking away. They'll quickly give up on a hunt for Basim. They'll see a cohort being yanked around a corner and think nothing of it. That breaks the immersion a bit, but it does make it easier to mess with these idiots.

I took some delight in tormenting my opponents, even if that may not match up to the code of conduct the assassins live by. One larger grunt was trapped in a room alone to guard a chest. I entered, used a smoke bomb to distract him, opened the chest and left, blocking the path behind me. I then made my way around to a gate that kept the guard locked in from the other side and spent a few minutes whistling at him, for no reason other than to annoy him and amuse myself.

The real star of the show is the version of ninth-century Baghdad Ubisoft has built. It feels rich and lived-in, with bystanders simply going about their day as a hooded figure darts by them to climb up the side of a building. Unfortunately, that level of detail wasn't reflected in the character models. Main characters and NPCs alike looked far less refined than their surroundings.

A hooded figure perched on a viewpoint looks toward a large green palace in a ninth-century version of Baghdad in Assassin's Creed Mirage.
Ubisoft

Some Arab critics and reviewers appreciated how Ubisoft represented Baghdad and Muslim culture in the game, and that's a positive sign. In that sense, Mirage seems like a prime candidate for the historical educational modes that Ubisoft has added to recent Assassin's Creed games.

I can't personally speak to the authenticity of the environment Ubisoft has created. The same goes for the Arabic used in the game, but the developers at least strove to avoid anachronisms. I spent an hour or so playing in Arabic with English subtitles and found it a compelling way to experience the game, though I missed hearing the velvet-voiced Shohreh Aghdashloo's portrayal of Basim's mentor Roshan too much.

Aghdashloo's performance is one of several highlights of a solid game. Developer Ubisoft Bordeaux has achieved what it set out to do in bringing back the format of early Assassin's Creed titles while adding some modern bells and whistles (such as a gameplay option to avoid the turgid pickpocketing minigame) and avoiding some of the old trappings.

No part of the game that I've encountered is set in the modern day. That's a wise move, since those parts of previous games pulled me out of the main experience and into some tedious sections that sought to serve a larger story. I didn't hear the word "animus" once this time around. Mirage does tie back into the broader Assassin's Creed narrative — Basim makes an appearance in Valhalla, after all — but you won't get sidetracked by Desmond Miles or Layla Hassan. That meant I could spend more of my time roaming the streets and rooftops of this well-crafted city, scouting enemy camps from above and figuring out the best way to approach an assassination mission.

Mirage probably won't be for everyone, including those who appreciated the format of the last three big Assassin's Creed games, but it struck a chord with me. Even though I've wrapped up the main story and have a bunch of other games to play (I'm looking at you, Cocoon and Spider-Man 2), I'll probably spend a little while longer nuzzled up in the comfort of Mirage.

Assassin's Creed Mirage is out now on PC, PlayStation 4, PlayStation 5, Xbox One and Xbox Series X/S. It's coming to iPhone 15 Pro devices next year.

This article originally appeared on Engadget at https://www.engadget.com/assassins-creed-mirage-review-a-warm-bloody-hug-from-an-old-friend-181918323.html?src=rss

ElevenLabs is building a universal AI dubbing machine

Posted on by

After Disney releases a new film in English, the company will go back and localize it in as many as 46 global languages to make the movie accesible to as wide an audience as possible. This is a massive undertaking, one for which Disney has an entire division — Disney Character Voices International Inc — to handle the task. And it's not like you're getting Chris Pratt back in the recording booth to dub his GotG III lines in Icelandic and Swahili — each version sounds a little different given the local voice actors. But with a new "AI dubbing" system from ElevenLabs, we could soon get a close recreation of Pratt's voice, regardless of the language spoken on-screen.   

ElevenLabs is an AI startup that offers a voice cloning service, allowing subscribers to generate nearly identical vocalizations with AI based on a few minutes worth of audio sample uploads. Not wholly unsurprising, as soon as the feature was released in beta, it was immediately exploited to impersonate celebrities, sometimes even without their prior knowledge and consent. 

The new AI dubbing feature does essentially the same thing — in more than 20 different languages including Hindi, Portuguese, Spanish, Japanese, Ukrainian, Polish and Arabic — but legitimately, and with permission. This tool is designed for use by media companies, educators and internet influencers who don't have Disney Money™ to fund their global adaptation efforts.

ElevenLabs asserts that the system will be able to not only translate "spoken content to another language in minutes" but also generate new spoken dialog in the target language using the actor's own voice. Or, at least, a AI generated recreation. The system is even reportedly capable of maintaining the "emotion and intonation" of the existing dialog and transferring that over to the generated translation.

 "It will help audiences enjoy any content they want, regardless of the language they speak," ElevenLabs CEO Mati Staniszewski said in a press statement. "And it will mean content creators can easily and authentically access a far bigger audience across the world."

This article originally appeared on Engadget at https://www.engadget.com/elevenlabs-is-building-a-universal-ai-dubbing-machine-130053504.html?src=rss

Amazon Alexa is evolving into a chatbot for your home

Posted on by

Amazon's Alexa is set to receive a major upgrade that will bring its conversational capabilities more in line with modern chatbots like Google Bard or OpenAI's ChatGPT, Dave Limp, SVP of Amazon Devices & Services, announced during the company's 2023 Devices event on Wednesday. The long-running digital assistant will soon be driven by a purpose-built large language model that will be available in nearly every new Echo device. 

“Our latest model has been specifically optimized for voice," Limp told the assembled crowd, "and the things we know our customers love — like having access to real-time information, efficiently controlling their smart home, and getting the most out of their home entertainment.”

Amazon is itself no stranger to genAI technology, having spent more than a decade researching its "ambient intelligence" systems. Generative AI models, specifically Alexa Teacher, have long driven the background functions of Alexa devices. "With generative AI within reach, we started doubling down on the home about nine years ago, and we had an epiphany, " Limp said. "We realized that all the investments in the R&D in the consumer electronics industry was being funneled into mobile phones. The SOCs, the displays, the chip sets, the sensors — it was be optimized to the phone." 

"That was understandable," he conceded. "It's a multi-billion-dollar-a-year industry. But at the same time, the place where you spend the vast majority of your life — your home — was virtually forgotten."

The new model will be both "larger and more generalized," Limp said, and will "help us take the next steps towards a remarkably different customer experience." To that end, Amazon set out to design the LLM based on five foundational capabilities and then tune the model specifically for voice applications rather than mobile screens. 

1. Conversational: We’ve studied what it takes to make a great conversation over the past nine years. It’s not just words; it’s body language, it’s understanding who you’re addressing, it’s eye contact and gestures.

2. Real-world applications: Alexa lives in the real world, not in the tab of your browser. And one of the unsolved challenges of these LLMs is how they interact with APIs and do the right thing.

3. Personalization: LLM in the home has to be personalized to you and your family.

4. Personality: “We’ve always said that the most boring dinner party is one where nobody has any opinions, and Alexa, powered by this LLM, will have opinions—and it will definitely still have the jokes and Easter eggs you’ve come to love from Alexa.”

5. Trust: To build an AI that will fulfill its promise, we need both trustworthiness and performance. “I have one of the most Alexa-fied houses out there, and I would not bring anything into my home that I felt compromised my family’s privacy.”

The voice optimizations simply mean that you won't have to repeat Alexa every time you talk to it. Customers enrolled in the company's Visual ID system will just need to face the screen before they start talking. What's more, the new Alexa will be more forgiving of stumbling or pause-filled speech, and it will soon modulate its tone and emotion based on the context of the conversation. 

The LLM will also be "connected to hundreds of thousands of real-world devices and services via APIs," the company's release reads. "It also enhances Alexa’s ability to process nuance and ambiguity—much like a person would—and intelligently take action." As such, users will soon be able to program complex requests, like “Alexa, every weeknight at 9 PM, make an announcement that it’s bedtime for the kids, dim the lights upstairs, turn on the porch light, and switch on the fan in the bedroom," all using just spoken commands. 

Limp tried to show off those natural conversation capabilities during an on-stage demonstration Wednesday, however the Alexa was not particularly cooperative, patently ignoring two of Limp's spoken prompts which required him to sheepishly repeat himself.    

The new model is far from Amazon's only genAI project. The company recently released a generative model to help its e-commerce sellers write product listings as well as incorporated a slew of AI-based features into its Thursday Night Football broadcasts at the start of the NFL season. The company has also weathered criticism from the Writers Guild of America over the retailer's allowance of AI-generated book listings which infringe heavily upon copyrighted works (and occasionally recommend eating suspect mushrooms). 

The new LLM will be available to existing Echo owners as part of a free preview on the devices they already own as well on every new Echo device sold, starting in 2024.

Follow all of the news live from Amazon’s 2023 Devices event right here.

This article originally appeared on Engadget at https://www.engadget.com/amazon-alexa-is-evolving-into-a-chatbot-for-your-home-152654742.html?src=rss

Nintendo’s new mobile game lets you pluck Pikmin on your browser

Posted on by

Nintendo has teamed up with Niantic for a new Pikmin mobile game that's mostly good for passing time than serious gaming. It's called Pikmin Finder, and as Nintendo Life notes, the companies have released it in time for the Nintendo Live event in Seattle. You can access the augmented reality game from any browser on your mobile, whether it's an iPhone or an Android device. We've tried it on several browsers, including Chrome and Opera, and we can verify that it works, as long as you allow it to access your camera. 

Similar to Pikmin Bloom, the game superimposes Pikmin on your environment as seen through your phone's camera. You can then pluck the creatures by swiping up — take note that there are typically more of the same color lurking around when you do spot one. Afterward, you can use the Pikmin you've plucked to search for treasures, including cakes and rubber duckies. You'll even see them bring you those treasures on your screen. 

Red pikmin superimposed on a keyboard.
Pikmin Finder

To play the game, you can go to its website on a mobile browser and start catching Pikmin on your phone. You can also scan the QR code that shows up on the website when you open it on a desktop browser.

This article originally appeared on Engadget at https://www.engadget.com/nintendos-new-mobile-game-lets-you-pluck-pikmin-on-your-browser-064423362.html?src=rss

New AP guidelines lay the groundwork for AI-assisted newsrooms

Posted on by

The Associated Press published standards today for generative AI use in its newsroom. The organization, which has a licensing agreement with ChatGPT maker OpenAI, listed a fairly restrictive and common-sense list of measures around the burgeoning tech while cautioning its staff not to use AI to make publishable content. Although nothing in the new guidelines is particularly controversial, less scrupulous outlets could view the AP’s blessing as a license to use generative AI more excessively or underhandedly.

The organization’s AI manifesto underscores a belief that artificial intelligence content should be treated as the flawed tool that it is — not a replacement for trained writers, editors and reporters exercising their best judgment. “We do not see AI as a replacement of journalists in any way,” the AP’s Vice President for Standards and Inclusion, Amanda Barrett, wrote in an article about its approach to AI today. “It is the responsibility of AP journalists to be accountable for the accuracy and fairness of the information we share.”

The article directs its journalists to view AI-generated content as “unvetted source material,” to which editorial staff “must apply their editorial judgment and AP’s sourcing standards when considering any information for publication.” It says employees may “experiment with ChatGPT with caution” but not create publishable content with it. That includes images, too. “In accordance with our standards, we do not alter any elements of our photos, video or audio,” it states. “Therefore, we do not allow the use of generative AI to add or subtract any elements.” However, it carved an exception for stories where AI illustrations or art are a story’s subject — and even then, it has to be clearly labeled as such.

Barrett warns about AI’s potential for spreading misinformation. To prevent the accidental publishing of anything AI-created that appears authentic, she says AP journalists “should exercise the same caution and skepticism they would normally, including trying to identify the source of the original content, doing a reverse image search to help verify an image’s origin, and checking for reports with similar content from trusted media.” To protect privacy, the guidelines also prohibit writers from entering “confidential or sensitive information into AI tools.”

Although that’s a relatively common-sense and uncontroversial set of rules, other media outlets have been less discerning. CNET was caught early this year publishing error-ridden AI-generated financial explainer articles (only labeled as computer-made if you clicked on the article’s byline). Gizmodo found itself in a similar spotlight this summer when it ran a Star Wars article full of inaccuracies. It’s not hard to imagine other outlets — desperate for an edge in the highly competitive media landscape — viewing the AP’s (tightly restricted) AI use as a green light to make robot journalism a central figure in their newsrooms, publishing poorly edited / inaccurate content or failing to label AI-generated work as such.

This article originally appeared on Engadget at https://www.engadget.com/new-ap-guidelines-lay-the-groundwork-for-ai-assisted-newsrooms-201009363.html?src=rss

Why humans can’t use natural language processing to speak with the animals

Posted on by

We’ve been wondering what goes on inside the minds of animals since antiquity. Dr. Doolittle’s talent was far from novel when it was first published in 1920; Greco-Roman literature is lousy with speaking animals, writers in Zhanguo-era China routinely ascribed language to certain animal species and they’re also prevalent in Indian, Egyptian, Hebrew and Native American storytelling traditions.

Even today, popular Western culture toys with the idea of talking animals, though often through a lens of technology-empowered speech rather than supernatural force. The dolphins from both Seaquest DSV and Johnny Mnemonic communicated with their bipedal contemporaries through advanced translation devices, as did Dug the dog from Up.

We’ve already got machine-learning systems and natural language processors that can translate human speech into any number of existing languages, and adapting that process to convert animal calls into human-interpretable signals doesn’t seem that big of a stretch. However, it turns out we’ve got more work to do before we can converse with nature.

What is language?

“All living things communicate,” an interdisciplinary team of researchers argued in 2018’s On understanding the nature and evolution of social cognition: a need for the study of communication. “Communication involves an action or characteristic of one individual that influences the behavior, behavioral tendency or physiology of at least one other individual in a fashion typically adaptive to both.”

From microbes, fungi and plants on up the evolutionary ladder, science has yet to find an organism that exists in such extreme isolation as to not have a natural means of communicating with the world around it. But we should be clear that “communication” and “language” are two very different things.

“No other natural communication system is like human language,” argues the Linguistics Society of America. Language allows us to express our inner thoughts and convey information, as well as request or even demand it. “Unlike any other animal communication system, it contains an expression for negation — what is not the case … Animal communication systems, in contrast, typically have at most a few dozen distinct calls, and they are used only to communicate immediate issues such as food, danger, threat, or reconciliation.”

That’s not to say that pets don’t understand us. “We know that dogs and cats can respond accurately to a wide range of human words when they have prior experience with those words and relevant outcomes,” Dr. Monique Udell, Director of the Human-Animal Interaction Laboratory at Oregon State University, told Engadget. “In many cases these associations are learned through basic conditioning,” Dr. Udell said — like when we yell “dinner” just before setting out bowls of food.

Whether or not our dogs and cats actually understand what “dinner” means outside of the immediate Pavlovian response — remains to be seen. “We know that at least some dogs have been able to learn to respond to over 1,000 human words (labels for objects) with high levels of accuracy,” Dr. Udell said. “Dogs currently hold the record among non-human animal species for being able to match spoken human words to objects or actions reliably,” but it’s “difficult to know for sure to what extent dogs understand the intent behind our words or actions.”

Dr. Udell continued: “This is because when we measure a dog or cat’s understanding of a stimulus, like a word, we typically do so based on their behavior.” You can teach a dog to sit with both English and German commands, but “if a dog responds the same way to the word ‘sit’ in English and in German, it is likely the simplest explanation — with the fewest assumptions — is that they have learned that when they sit in the presence of either word then there is a pleasant consequence.”

Tea Stražičić for Engadget/Silica Magazine

Hush, the computers are speaking

Natural Language Programming (NLP) is the branch of AI that enables computers and algorithmic models to interpret text and speech, including the speaker’s intent, the same way we meatsacks do. It combines computational linguistics, which models the syntax, grammar and structure of a language, and machine-learning models, which “automatically extract, classify, and label elements of text and voice data and then assign a statistical likelihood to each possible meaning of those elements,” according to IBM. NLP underpins the functionality of every digital assistant on the market. Basically any time you’re speaking at a “smart” device, NLP is translating your words into machine-understandable signals and vice versa.

The field of NLP research has undergone a significant evolution in recent years, as its core systems have migrated from older Recurrent and Convoluted Neural Networks towards Google’s Transformer architecture, which greatly increases training efficiency.

Dr. Noah D. Goodman, Associate Professor of Psychology and Computer Science, and Linguistics at Stanford University, told Engadget that, with RNNs, “you'll have to go time-step by time-step or like word by word through the data and then do the same thing backward.” In contrast, with a transformer, “you basically take the whole string of words and push them through the network at the same time.”

“It really matters to make that training more efficient,” Dr. Goodman continued. “Transformers, they're cool … but by far the biggest thing is that they make it possible to train efficiently and therefore train much bigger models on much more data.”

Talkin’ jive ain’t just for turkeys

While many species’ communication systems have been studied in recent years — most notably cetaceans like whales and dolphins, but also the southern pied babbler, for its song’s potentially syntactic qualities, and vervet monkeys’ communal predator warning system — none have shown the sheer degree of complexity as the call of the avian family Paridae: the chickadees, tits and titmice.

Dr. Jeffrey Lucas, professor in the Biological Sciences department at Purdue University, told Engadget that the Paridae call “is one of the most complicated vocal systems that we know of. At the end of the day, what the [field’s voluminous number of research] papers are showing is that it's god-awfully complicated, and the problem with the papers is that they grossly under-interpret how complicated [the calls] actually are.”

These parids often live in socially complex, heterospecific flocks, mixed groupings that include multiple songbird and woodpecker species. The complexity of the birds’ social system is correlated with an increased diversity in communications systems, Dr. Lucas said. “Part of the reason why that correlation exists is because, if you have a complex social system that's multi-dimensional, then you have to convey a variety of different kinds of information across different contexts. In the bird world, they have to defend their territory, talk about food, integrate into the social system [and resolve] mating issues.”

The chickadee call consist of at least six distinct notes set in an open-ended vocal structure, which is both monumentally rare in non-human communication systems and the reason for the Chickadee’s call complexity. An open-ended vocal system means that “increased recording of chick-a-dee calls will continually reveal calls with distinct note-type compositions,” explained the 2012 study, Linking social complexity and vocal complexity: a parid perspective. “This open-ended nature is one of the main features the chick-a-dee call shares with human language, and one of the main differences between the chick-a-dee call and the finite song repertoires of most songbird species.”

Dolphin translation by Tea Stražičić
Tea Stražičić for Engadget/Silica Magazine

Dolphins have no need for kings

Training language models isn’t simply a matter of shoving in large amounts of data. When training a model to translate an unknown language into what you’re speaking, you need to have at least a rudimentary understanding of how the the two languages correlate with one another so that the translated text retains the proper intent of the speaker.

“The strongest kind of data that we could have is what's called a parallel corpus,” Dr. Goodman explained, which is basically having a Rosetta Stone for the two tongues. In that case, you’d simply have to map between specific words, symbols and phonemes in each language — figure out what means “river” or “one bushel of wheat” in each and build out from there.

Without that perfect translation artifact, so long as you have large corpuses of data for both languages, “it's still possible to learn a translation between the languages, but it hinges pretty crucially on the idea that the kind of latent conceptual structure,” Dr. Goodman continued, which assumes that both culture’s definitions of “one bushel of wheat” are generally equivalent.

Goodman points to the word pairs ’man and woman’ and ’king and queen’ in English. “The structure, or geometry, of that relationship we expect English, if we were translating into Hungarian, we would also expect those four concepts to stand in a similar relationship,” Dr. Goodman said. “Then effectively the way we'll learn a translation now is by learning to translate in a way that preserves the structure of that conceptual space as much as possible.”

Having a large corpus of data to work with in this situation also enables unsupervised learning techniques to be used to “extract the latent conceptual space,” Dr. Goodman said, though that method is more resource intensive and less efficient. However, if all you have is a large corpus in only one of the languages, you’re generally out of luck.

“For most human languages we assume the [quartet concepts] are kind of, sort of similar, like, maybe they don't have ‘king and queen’ but they definitely have ‘man and woman,’” Dr. Goodman continued. ”But I think for animal communication, we can't assume that dolphins have a concept of ‘king and queen’ or whether they have ‘men and women.’ I don't know, maybe, maybe not.”

And without even that rudimentary conceptual alignment to work from, discerning the context and intent of a animal’s call — much less, deciphering the syntax, grammar and semantics of the underlying communication system — becomes much more difficult. “You're in a much weaker position,” Dr. Goodman said. “If you have the utterances in the world context that they're uttered in, then you might be able to get somewhere.”

Basically, if you can obtain multimodal data that provides context for the recorded animal call — the environmental conditions, time of day or year, the presence of prey or predator species, etc — you can “ground” the language data into the physical environment. From there you can “assume that English grounds into the physical environment in the same way as this weird new language grounds into the physical environment’ and use that as a kind of bridge between the languages.”

Unfortunately, the challenge of translating bird calls into English (or any other human language) is going to fall squarely into the fourth category. This means we’ll need more data and a lot of different types of data as we continue to build our basic understanding of the structures of these calls from the ground up. Some of those efforts are already underway.

The Dolphin Communication Project, for example, employs a combination “mobile video/acoustic system” to capture both the utterances of wild dolphins and their relative position in physical space at that time to give researchers added context to the calls. Biologging tags — animal-borne sensors affixed to hide, hair, or horn that track the locations and conditions of their hosts — continue to shrink in size while growing in both capacity and capability, which should help researchers gather even more data about these communities.

What if birds are just constantly screaming about the heat?

Even if we won’t be able to immediately chat with our furred and feathered neighbors, gaining a better understanding of how they at least talk to each other could prove valuable to conservation efforts. Dr. Lucas points to a recent study he participated in that found environmental changes induced by climate change can radically change how different bird species interact in mixed flocks. “What we showed was that if you look across the disturbance gradients, then everything changes,” Dr. Lucas said. “What they do with space changes, how they interact with other birds changes. Their vocal systems change.”

“The social interactions for birds in winter are extraordinarily important because you know, 10 gram bird — if it doesn't eat in a day, it's dead,” Dr. Lucas continued. “So information about their environment is extraordinarily important. And what those mixed species flocks do is to provide some of that information.”

However that network quickly breaks down as the habitat degrades and in order to survive “they have to really go through fairly extreme changes in behavior and social systems and vocal systems … but that impacts fertility rates, and their ability to feed their kids and that sort of thing.”

Better understanding their calls will help us better understand their levels of stress, which can serve both modern conservation efforts and agricultural ends. “The idea is that we can get an idea about the level of stress in [farm animals], then use that as an index of what's happening in the barn and whether we can maybe even mitigate that using vocalizations,” Dr. Lucas said. “AI probably is going to help us do this.”

“Scientific sources indicate that noise in farm animal environments is a detrimental factor to animal health,” Jan Brouček of the Research Institute for Animal Production Nitra, observed in 2014. “Especially longer lasting sounds can affect the health of animals. Noise directly affects reproductive physiology or energy consumption.” That continuous drone is thought to also indirectly impact other behaviors including habitat use, courtship, mating, reproduction and the care of offspring. 

Conversely, 2021’s research, The effect of music on livestock: cattle, poultry and pigs, has shown that playing music helps to calm livestock and reduce stress during times of intensive production. We can measure that reduction in stress based on what sorts of happy sounds those animals make. Like listening to music in another language, we can get with the vibe, even if we can't understand the lyrics

This article originally appeared on Engadget at https://www.engadget.com/why-humans-cant-use-natural-language-processing-to-speak-with-the-animals-143050169.html?src=rss

Hitting the Books: The dangerous real-world consequences of our online attention economy

Posted on by

If reality television has taught us anything, it's there's not much people won't do if offered enough money and attention. Sometimes, even just the latter. Unfortunately for the future prospects of our civilization, modern social media has focused upon those same character foibles and optimized them at a global scale, sacrifices at the altar of audience growth and engagement. In Outrage Machine, writer and technologist Tobias Rose-Stockwell, walks readers through the inner workings of these modern technologies, illustrating how they're designed to capture and keep our attention, regardless of what they have to do in order to do it. In the excerpt below, Rose-Stockwell examines the human cost of feeding the content machine through a discussion on YouTube personality Nikocado Avocado's rise to internet stardom.

 

lots of angry faces, black text white background
Legacy Lit

Excerpted from OUTRAGE MACHINE: How Tech Amplifies Discontent, Disrupts Democracy—And What We Can Do About It by Tobias Rose-Stockwell. Copyright © 2023 by Tobias Rose-Stockwell. Reprinted with permission of Legacy Lit. All rights reserved.

This Game Is Not Just a Game

Social media can seem like a game. When we open our apps and craft a post, the way we look to score points in the form of likes and followers distinctly resembles a strange new playful competition. But while it feels like a game, it is unlike any other game we might play in our spare time.

The academic C. Thi Nguyen has explained how games are different: “Actions in games are screened off, in important ways, from ordinary life. When we are playing basketball, and you block my pass, I do not take this to be a sign of your long-term hostility towards me. When we are playing at having an insult contest, we don’t take each other’s speech to be indicative of our actual attitudes or beliefs about the world.” Games happen in what the Dutch historian Johan Huizinga famously called “the magic circle”— where the players take on alternate roles, and our actions take on alternate meanings.

With social media we never exit the game. Our phones are always with us. We don’t extricate ourselves from the mechanics. And since the goal of the game designers of social media is to keep us there as long as possible, it’s an active competition with real life. With a constant type of habituated attention being pulled into the metrics, we never leave these digital spaces. In doing so, social media has colonized our world with its game mechanics.

Metrics are Money

While we are paid in the small rushes of dopamine that come from accumulating abstract numbers, metrics also translate into hard cash. Acquiring these metrics don’t just provide us with hits of emotional validation. They are transferable into economic value that is quantifiable and very real.

It’s no secret that the ability to consistently capture attention is an asset that brands will pay for. A follower is a tangible, monetizable asset worth money. If you’re trying to purchase followers, Twitter will charge you between $2 and $4 to acquire a new one using their promoted accounts feature.

If you have a significant enough following, brands will pay you to post sponsored items on their behalf. Depending on the size of your following in Instagram, for instance, these payouts can range from $75 per post (to an account with two thousand followers), up to hundreds of thousands of dollars per post (for accounts with hundreds of thousands of followers).

Between 2017 and 2021, the average cost for reaching a thousand Twitter users (the metric advertisers use is CPM, or cost per mille) was between $5 and $7. It costs that much to get a thousand eyeballs on your post. Any strategies that increase how much your content is shared also have a financial value.

Let’s now bring this economic incentive back to Billy Brady’s accounting of the engagement value of moral outrage. He found that adding a single moral or emotional word to a post on Twitter increased the viral spread of that content by 17 percent per word. All of our posts to social media exist in a marketplace for attention — they vie for the top of our followers’ feeds. Our posts are always competing against other people’s posts. If outraged posts have an advantage in this competition, they are literally worth more money.

For a brand or an individual, if you want to increase the value of a post, then including moral outrage, or linking to a larger movement that signals its moral conviction, might increase the reach of that content by at least that much. Moreover, it might actually improve the perception and brand affinity by appealing to the moral foundations of the brand’s consumers and employees, increasing sales and burnishing their reputation. This can be an inherently polarizing strategy, as a company that picks a cause to support, whose audience is morally diverse, might then alienate a sizable percentage of their customer base who disagree with that cause. But these economics can also make sense — if a company knows enough about its consumers’ and employees’ moral affiliations — it can make sure to pick a cause-sector that’s in line with its customers.

Since moral content is a reliable tool for capturing attention, it can also be used for psychographic profiling for future marketing opportunities. Many major brands do this with tremendous success — creating viral campaigns that utilize moral righteousness and outrage to gain traction and attention among core consumers who have a similar moral disposition. These campaigns also often get a secondary boost due to the proliferation of pile- ons and think pieces discussing these ad spots. Brands that moralize their products often succeed in the attention marketplace.

This basic economic incentive can help to explain how and why so many brands have begun to link themselves with online cause-related issues. While it may make strong moral sense to those decision-makers, it can make clear economic sense to the company as a whole as well. Social media provides measurable financial incentives for companies to include moral language in their quest to burnish their brands and perceptions.

But as nefarious as this sounds, moralization of content is not always the result of callous manipulation and greed. Social metrics do something else that influences our behavior in pernicious ways.

Audience Capture

In the latter days of 2016, I wrote an article about how social media was diminishing our capacity for empathy. In the wake of that year’s presidential election, the article went hugely viral, and was shared with several million people. At the time I was working on other projects full time. When the article took off, I shifted my focus away from the consulting work I had been doing for years, and began focusing instead on writing full time. One of the by-products of that tremendous signal from this new audience is the book you’re reading right now.

A sizable new audience of strangers had given me a clear message: This was important. Do more of it. When many people we care about tell us what we should be doing, we listen.

This is the result of “audience capture”: how we influence, and are influenced by those who observe us. We don’t just capture an audience — we are also captured by their feedback. This is often a wonderful thing, provoking us to produce more useful and interesting works. As creators, the signal from our audience is a huge part of why we do what we do.

But it also has a dark side. The writer Gurwinder Boghal has explained the phenomena of audience capture for influencers illustrating the story of a young YouTuber named Nicholas Perry. In 2016, Perry began a You- Tube channel as a skinny vegan violinist. After a year of getting little traction online, he abandoned veganism, citing health concerns, and shifted to uploading mukbang (eating show) videos of him trying different foods for his followers. These followers began demanding more and more extreme feats of food consumption. Before long, in an attempt to appease his increasingly demanding audience, he was posting videos of himself eating whole fast-food menus in a single sitting.

He found a large audience with this new format. In terms of metrics, this new format was overwhelmingly successful. After several years of following his audience’s continued requests, he amassed millions of followers, and over a billion total views. But in the process, his online identity and physical character changed dramatically as well. Nicholas Perry became the personality Nikocado — an obese parody of himself, ballooning to more than four hundred pounds, voraciously consuming anything his audience asked him to eat. Following his audience’s desires caused him to pursue increasingly extreme feats at the expense of his mental and physical health.

a horrifying before and after
Legacy Lit

Nicholas Perry, left, and Nikocado, right, after several years of building a following on YouTube. Source: Nikocado Avocado YouTube Channel.

Boghal summarizes this cross-directional influence.

When influencers are analyzing audience feedback, they often find that their more outlandish behavior receives the most attention and approval, which leads them to recalibrate their personalities according to far more extreme social cues than those they’d receive in real life. In doing this they exaggerate the more idiosyncratic facets of their personalities, becoming crude caricatures of themselves.

This need not only apply to influencers. We are signal-processing machines. We respond to the types of positive signals we receive from those who observe us. Our audiences online reflect back to us what their opinion of our behavior is, and we adapt to fit it. The metrics (likes, followers, shares, and comments) available to us now on social media allow for us to measure that feedback far more precisely than we previously could, leading to us internalizing what is “good” behavior.

As we find ourselves more and more inside of these online spaces, this influence becomes more pronounced. As Boghal notes, “We are all gaining online audiences.” Anytime we post to our followers, we are entering into a process of exchange with our viewers — one that is beholden to the same extreme engagement problems found everywhere else on social media.

This article originally appeared on Engadget at https://www.engadget.com/hitting-the-books-the-dangerous-real-world-consequences-of-our-online-attention-economy-143050602.html?src=rss

Tor’s shadowy reputation will only end if we all use it

Posted on by

“Tor” evokes an image of the dark web; a place to hire hitmen or buy drugs that, at this point, is overrun by feds trying to catch you in the act. The reality, however, is a lot more boring than that — but it’s also more secure.

The Onion Router, now called Tor, is a privacy-focused web browser run by a nonprofit group. You can download it for free and use it to shop online or browse social media, just like you would on Chrome or Firefox or Safari, but with additional access to unlisted websites ending in .onion. This is what people think of as the “dark web,” because the sites aren’t indexed by search engines. But those sites aren’t an inherently criminal endeavor.

“This is not a hacker tool,” said Pavel Zoneff, director of strategic communications at The Tor Project. “It is a browser just as easy to use as any other browser that people are used to.”

That’s right, despite common misconceptions, Tor can be used for any internet browsing you usually do. The key difference with Tor is that the network hides your IP address and other system information for full anonymity. This may sound familiar, because it’s how a lot of people approach VPNs, but the difference is in the details.

VPNs are just encrypted tunnels hiding your traffic from one hop to another. The company behind a VPN can still access your information, sell it or pass it along to law enforcement. With Tor, there’s no link between you and your traffic, according to Jed Crandall, an associate professor at Arizona State University. Tor is built in the “higher layers” of the network and routes your traffic through separate tunnels, instead of a single encrypted tunnel. While the first tunnel may know some personal information and the last one may know the sites you visited, there is virtually nothing connecting those data points because your IP address and other identifying information are bounced from server to server into obscurity.

In simpler terms: using regular browsers directly connects you and your traffic, adding a VPN routes that information through an encrypted tunnel so that your internet service provider can’t see it and Tor scatters your identity and your search traffic until it becomes almost anonymous, and very difficult to identify.

Accessing unindexed websites adds extra perks, like secure communication. While a platform like WhatsApp offers encrypted conversations, there could be traces that the conversation happened left on the device if it’s ever investigated, according to Crandall. Tor's communication tunnels are secure and much harder to trace that the conversation ever happened.

Other use cases may include keeping the identities of sensitive populations like undocumented immigrants anonymous, trying to unionize a workplace without the company shutting it down, victims of domestic violence looking for resources without their abuser finding out or, as Crandall said, wanting to make embarrassing Google searches without related targeted ads following you around forever.

Still, with added layers of security can come some additional hiccups, like lag or longer loading times. That could be true for some users depending on what they do online, but anecdotally it's gotten a lot faster in recent years, and users have said they barely notice a difference compared to other browsers. Sameer Patil, associate professor at the School of Computing at the University of Utah, studied this by having students and staff try out Tor as their main browser. “I was personally very surprised at how many sites and things just work fine in the Tor browser. So not only did they work as intended, but they also were fast enough,” Patil said.

But even if online privacy isn’t your main concern personally, using Tor can help support industries that heavily rely on it. By using the anonymous and secure browser, you’re supporting activists, journalists and everyone else’s privacy because the more people that use it, the more secure it gets, according to Patil. If only certain sensitive groups use it, it’ll be easier to deanonymize and ultimately track down identities. When you’re one in a billion using it, that task becomes nearly impossible.

This article originally appeared on Engadget at https://www.engadget.com/tor-dark-web-privacy-secure-browser-anonymous-130048839.html?src=rss

Astrophysicist who claimed to find alien tech may have done the science wrong

Posted on by

Last month, theoretical physicist Avi Loeb made headlines with the sensational claim that tiny spherules recovered from the bottom of the ocean were probably of alien origin. “It’s most likely a technological gadget with artificial intelligence,” he said to The New York Times, which published a story today about the Harvard professor’s contentious claims. Although the biggest scientific breakthroughs often start with a bold hypothesis, Loeb’s peers believe the decorated astrophysicist’s assertions can be called many things — but “good science” isn’t one of them.

Loeb’s proclamations stem from an object that US government sensors logged on January 8th, 2014: a fireball from space that blazed into the western Pacific Ocean off the northeastern coast of Papua New Guinea. Highlighting its logged speed and direction as an anomaly, Loeb and undergraduate assistant Amir Siraj targeted the otherwise inconsequential planetary entry as an object worthy of further investigation.

Fast-forward to last month, when Loeb led a voyage — funded by a crypto entrepreneur — to recover evidence from the fireball’s calculated crash path. Dragging a magnetic sled attached to the expedition boat across the ocean floor, the team recovered a series of tiny spherical objects which Loeb says “appear under a microscope as beautiful metallic marbles.” Preliminary analysis indicated that the sub-millimeter orbs were 84 percent iron, with silicon, magnesium and trace elements comprising the rest. Loeb believes that “as a result of being exposed to the fireball’s heat, the surface of the object likely disintegrated into tiny spherules, similar in number per unit area to those recovered by the expedition.”

An electron microprobe image of one of the spherules recovered from the ocean's bottom.
Avi Loeb / Medium

Not one to exercise much caution with public pronouncements, Loeb wrote in a Medium post, “Their discovery opens a new frontier in astronomy, where what lay outside the solar system is studied through a microscope rather than a telescope.” He summarized, in an equally dramatic manner, “The discovery of spherules felt like a miracle.” Soon after, CBS News picked up on his excitement and published an attention-grabbing article titled, “Harvard professor Avi Loeb believes he’s found fragments of alien technology.” Loeb has sent the mysterious spheres to Harvard University, the University of California, Berkeley and the Bruker Corporation in Germany for more in-depth analysis.

“It has material strength that is tougher than all space rock that were seen before, and catalogued by NASA,” CBS Newsreported Loeb as saying earlier this month. “We calculated its speed outside the solar system. It was 60 km per second, faster than 95% of all stars in the vicinity of the sun. The fact that it was made of materials tougher than even iron meteorites, and moving faster than 95% of all stars in the vicinity of the sun, suggested potentially it could be a spacecraft from another civilization or some technological gadget.”

It all sounds fascinating, especially with the resurgent interest in UFOs and the quest to discover signs of alien life. But there’s one problem: The scientific community, by and large, believes Loeb is, if not entirely full of it, practicing something far outside what they’d call science.

Peter Brown, a meteor physicist at Western University in Ontario, said that “several percent” of detected events appear interstellar at first but almost always end up chalked up to a measurement error. Steve Desch, an astrophysicist at Arizona State University, argued at a recent conference that if the object were traveling as fast as the data suggests — one of the points Loeb uses to indicate its origin was from outside our solar system — it would have been wholly incinerated entering the Earth’s atmosphere. Brown and other scientists also highlight Loeb’s lack of engagement with peers who study similar unidentified fireballs.

Brown recently presented data (accepted for publication in The Astrophysical Journal) demonstrating that NASA’s recordings in cases like these often end up being proven untrustworthy. He believes the fireball likely impacted at a slower speed than the recorded data suggested. “If the speed was overestimated, then the object becomes, more or less, within the realm of what we see in terms of other bound solar system objects,” he said. (Loeb retorted by citing an unbendable trust in government data: “They are responsible for national security. I think they know what they are doing.”) The New York Times adds that the government is unlikely to declassify the data that would allow the scientific community to learn how precise (or not) it is.

Harvard astrophysicist Avi Loeb holding a vial while looking at the camera with a proud smile.
Avi Loeb / Medium

Regardless of the spherules’ origins, researchers are alarmed by Loeb’s penchant for venturing outside science to make bold (and highly publicized) claims — with his scientific background boosting their perceived legitimacy. The gist of their alarm is that becoming a Harvard-employed astrophysicist doesn’t grant you the wizard-like ability to know the answers to questions the scientific method hasn’t yet confirmed. On the contrary, it’s supposed to mean your peers respect you for exercising restraint and doing quite the opposite. “[Loeb’s claims are] a real breakdown of the peer review process and the scientific method,” Desch said to The New York Times. “And it’s so demoralizing and tiring.”

Loeb’s views about his peers’ harsh response can be summarized in his cited quote from philosopher Arthur Schopenhauer from a recent blog post. “All truth passes through three stages: First, it is ridiculed; second, it is violently opposed; and third, it is accepted as self-evident.” Notably, Loeb seemingly refers to his conclusions about the preliminary findings — with plenty of question marks still intact — as “truth.”

The Oxford English Dictionary defines confirmation bias as “the tendency to interpret new evidence as confirmation of one's existing beliefs or theories.” Loeb’s words and excited tone suggest he knows the answer and that his peers’ criticism stems from their resistance to the new frontier he’s discovered. However, their criticism seems only partially about his specific conclusions; it’s paired with a larger concern about an esteemed cohort jumping to conclusions that fall far outside of the scientific method. “What the public is seeing in Loeb is not how science works,” remarked Desch. “And they shouldn’t go away thinking that.”

This article originally appeared on Engadget at https://www.engadget.com/astrophysicist-who-claimed-to-find-alien-tech-may-have-done-the-science-wrong-214008434.html?src=rss