Thousands of people took to the streets in Dublin to attend a Halloween parade that never came, according to reporting by The Independent. Why did they do such a thing? It was all due to an AI-generated ad that promoted the fake event.
The My Spirit Halloween website advertised the completely fabricated Macnas Halloween Parade, which was supposed to take place from 7PM to 9PM on the streets of Dublin. News of the parade quickly spread online, and it even received a news ranking on Google.
So, yeah, thousands of people put on their Halloween costumes and stood on the street for a couple of hours, waiting for an event that would never happen. They even left room in the street for the parade to pass by. That’s thoughtful, but also a bit sad.
The situation forced Ireland’s police force to put out a message to urge would-be parade-goers to “disperse safely.” A spokesperson for the organization said that “contrary to information being circulated online, no Halloween parade is scheduled to take place in Dublin city center this evening or tonight.”
Please be advised that contrary to information being circulated online, no Halloween parade is scheduled to take place in Dublin City Centre this evening or tonight.
All those gathered on O’Connell Street in expectation of such a parade are asked to disperse safely.
Local city councilor Janice Boylan expressed disappointment over the whole situation.“Everyone is trying to have a fun and safe Halloween. Having a parade to go to sounded really good,” she said. “I know an awful lot of people turned up. It’s a terrible pity.”
The My Spirit Halloween listing has been taken down, but there remains the question as to why it turned up in the first place. It’s worth noting that this is a different entity from the popular seasonal retailer Spirit Halloween.
The My Spirit Halloween website looks to come from Pakistan and posts all kinds of AI-generated content like the ad that caused all of this trouble, according to Yahoo News. This particular post happened to get picked up by TikTok and Google, causing the rapid dissemination of the, literal, fake news.
It’s pretty wild, right? An AI-generated post likely created in Pakistan caused thousands of actual people to take to the streets halfway across the globe. What is that curse again? Oh yeah. “May you live in interesting times.”
This article originally appeared on Engadget at https://www.engadget.com/ai/an-ai-generated-ad-left-thousands-of-dubliners-waiting-for-a-halloween-parade-that-never-came-162550781.html?src=rss
Google CEO Sundar Pichai just revealed that AI now generates more than a quarter of new code for its products, according to a company earnings call transcribed by Ars Technica. In other words, AI tools are already having an absolutely mammoth impact on the development of software.
Pichai did say that human programmers oversee the computer-generated code, which is something. The CEO noted that AI coding helps with “boosting productivity and efficiency," ensuring that engineers “do more and move faster.”
There’s no two ways around it. 25 percent is a lot, and Google is just one company relying on AI algorithms to perform complex coding tasks. According to Stack Overflow’s 2024 Developer Survey, over 75 percent of respondents are already using or are “planning to use” AI tools to assist with software development. Another survey by GitHub indicated that 92 percent of US-based developers are currently using AI coding tools.
This leads us to the rampaging elephant in the room. As AI continues to gobble up coding tasks, human experience starts to dwindle. This could eventually lead to a decreased knowledge base in which humans don’t know how to fix errors created by AI algorithms that were, in turn, created by other AI algorithms. We could be staring down an ouroboros of confusion where it’s nearly impossible to detect bugs amidst generations of AI code. Fun times!
We aren’t quite there yet, but AI-assisted coding shows no signs of slowing down. The process started its meteoric rise back in 2022 when GitHub widely launched its Copilot program. Since then, companies like Anthropic, Meta, Google and OpenAI have all released AI-coding software suites. GitHub recently announced that Copilot can now be used with models from Anthropic and Google, in addition to OpenAI.
This article originally appeared on Engadget at https://www.engadget.com/ai/google-ceo-says-a-quarter-of-the-companys-new-code-is-already-ai-generated-180038896.html?src=rss
In a blog post today, Google said it will make AI Overviews in Google Search available to users living in more than 100 countries. Google says people like AI Overviews, so the company plans to reach 1 billion users each month. Additionally, Google is expanding language support — now, users in any country where AI Overviews can get them in any available language. For example users in Japan are no longer limited to Japanese.
If you aren’t familiar with AI Overviews by now, it’s a feature found in Google Search. Google’s Gemini AI model will generate a summary for most things people search for, or at least when Google's systems detect a query where generative AI can be useful. This summary appears at the top of the page. Besides answering questions, it can also act as a meal planner, vacation organizer and brainstorming tool, among many other things.
This update follows several released over the last few months since the introduction of AI Overviews in May. The most recent update before this one allowed users to visit websites by clicking on links within the AI Overview text. Google says that this has resulted in a traffic increase to “supporting websites.” However, some publishers and sites fear that the overviews are destroying their traffic by being the first thing people see.
News like this isn’t surprising, as Google has been stuffing as much AI into Search as it can. AI Overviews have also generated some controversy after the AI gave people inaccurate information for searches — for example, it recommended putting glue on pizza for a bit. In response to those issues, Google put some protections and guardrails on AI overviews to avoid similar strange issues.
Correction, October 28 2024, 1:30PM ET: This story originally stated that Google was adding local language support to AI Overviews. That feature was already available, so we've clarified that any supported language can be displayed in any country where AI Overviews are available. We apologize for the error.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/google-is-expanding-ai-overviews-in-search-to-more-than-100-countries-170925933.html?src=rss
In a blog post today, Google said it will make AI Overviews in Google Search available to users living in more than 100 countries. Google says people like AI Overviews, so the company plans to reach 1 billion users each month. Additionally, Google is expanding language support — now, users in any country where AI Overviews can get them in any available language. For example users in Japan are no longer limited to Japanese.
If you aren’t familiar with AI Overviews by now, it’s a feature found in Google Search. Google’s Gemini AI model will generate a summary for most things people search for, or at least when Google's systems detect a query where generative AI can be useful. This summary appears at the top of the page. Besides answering questions, it can also act as a meal planner, vacation organizer and brainstorming tool, among many other things.
This update follows several released over the last few months since the introduction of AI Overviews in May. The most recent update before this one allowed users to visit websites by clicking on links within the AI Overview text. Google says that this has resulted in a traffic increase to “supporting websites.” However, some publishers and sites fear that the overviews are destroying their traffic by being the first thing people see.
News like this isn’t surprising, as Google has been stuffing as much AI into Search as it can. AI Overviews have also generated some controversy after the AI gave people inaccurate information for searches — for example, it recommended putting glue on pizza for a bit. In response to those issues, Google put some protections and guardrails on AI overviews to avoid similar strange issues.
Correction, October 28 2024, 1:30PM ET: This story originally stated that Google was adding local language support to AI Overviews. That feature was already available, so we've clarified that any supported language can be displayed in any country where AI Overviews are available. We apologize for the error.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/google-is-expanding-ai-overviews-in-search-to-more-than-100-countries-170925933.html?src=rss
Lyft has agreed to to tell its drivers how much they can truly earn on the ride-hailing platform — and back it up with evidence — as part of its settlement for a lawsuit filed by the US Justice Department and the Federal Trade Commission. The lawsuit accused the company of making "numerous false and misleading claims" in the advertisements it released in 2021 and 2022, when the demand for rides recovered following COVID-19 lockdowns in the previous years. Lyft promised drivers up to $43 an hour in some locations, the FTC said, without revealing that those numbers were based on the earnings of its top drivers.
The rates it published allegedly didn't represent drivers' average earnings and inflated actual earnings by up to 30 percent. Further, the FTC said that Lyft "failed to disclose" that information, as well as the fact that the amounts it published included passengers' tips. The company also promised in its ads that drivers will get paid a set amount if they complete a certain number of rides within a specific timeframe. A driver is supposed to make $975, for instance, if they complete 45 rides over a weekend.
Lyft allegedly didn't clarify that it will only pay the difference between the what the drivers' earn and its promised guaranteed earnings. Drivers thought they were getting those guaranteed payments on top of their ride payments as a bonus for completing a specific number of rides. The FTC accused Lyft of continuing to make "deceptive earnings claims" even after it sent the company a notice of its concerns in October 2021, as well.
Earlier this month, the company launched an earnings dashboard that showed the estimated hourly rate for each ride, along with the driver's daily, weekly and yearly earnings. But under the settlement, Lyft will have to explicitly tell drivers how much their potential take-home pay is based on typical, instead of inflated, earnings. It has to take tips out of the equation, and it has to to clarify that it will only pay the difference between what the drivers get from rides and its guaranteed earnings promise. Finally, it will have to pay a $2.1 million civil penalty.
This article originally appeared on Engadget at https://www.engadget.com/transportation/lyft-will-have-to-tell-drivers-how-much-they-can-truly-earn-with-evidence-120011572.html?src=rss
Lyft has agreed to to tell its drivers how much they can truly earn on the ride-hailing platform — and back it up with evidence — as part of its settlement for a lawsuit filed by the US Justice Department and the Federal Trade Commission. The lawsuit accused the company of making "numerous false and misleading claims" in the advertisements it released in 2021 and 2022, when the demand for rides recovered following COVID-19 lockdowns in the previous years. Lyft promised drivers up to $43 an hour in some locations, the FTC said, without revealing that those numbers were based on the earnings of its top drivers.
The rates it published allegedly didn't represent drivers' average earnings and inflated actual earnings by up to 30 percent. Further, the FTC said that Lyft "failed to disclose" that information, as well as the fact that the amounts it published included passengers' tips. The company also promised in its ads that drivers will get paid a set amount if they complete a certain number of rides within a specific timeframe. A driver is supposed to make $975, for instance, if they complete 45 rides over a weekend.
Lyft allegedly didn't clarify that it will only pay the difference between the what the drivers' earn and its promised guaranteed earnings. Drivers thought they were getting those guaranteed payments on top of their ride payments as a bonus for completing a specific number of rides. The FTC accused Lyft of continuing to make "deceptive earnings claims" even after it sent the company a notice of its concerns in October 2021, as well.
Earlier this month, the company launched an earnings dashboard that showed the estimated hourly rate for each ride, along with the driver's daily, weekly and yearly earnings. But under the settlement, Lyft will have to explicitly tell drivers how much their potential take-home pay is based on typical, instead of inflated, earnings. It has to take tips out of the equation, and it has to to clarify that it will only pay the difference between what the drivers get from rides and its guaranteed earnings promise. Finally, it will have to pay a $2.1 million civil penalty.
This article originally appeared on Engadget at https://www.engadget.com/transportation/lyft-will-have-to-tell-drivers-how-much-they-can-truly-earn-with-evidence-120011572.html?src=rss
Blumhouse wasn’t going to publish a game in 2024. The studio, one of the leading names in horror films, announced in February 2023 that it was launching a video game publishing business and executives were scouting projects from independent teams with budgets under $10 million. The goal of Blumhouse Games was to support a few rad horror titles per year, with a tentative plan to start publishing them in 2025.
But then, in September 2023, the Blumhouse folks stumbled across Fear the Spotlight. It was a moody, voxelized horror game about two friends sneaking around their haunted high school and communing with the ghosts of students that died in a fire in the ’90s. Spooky shenanigans and mysterious puzzles ensued, all presented in third-person and with a gritty PS One aesthetic. Fear the Spotlight was a PC game made by Crista Castro and Bryan Singh of Cozy Game Pals, a husband-and-wife team with impressive professional credentials: Castro was an art director at Nickelodeon and the art lead on the Animaniacs reboot, and Singh was a programmer who worked on The Last of Us and Uncharted series and Journey. They left their corporate jobs and founded Cozy Game Pals during the pandemic, and Fear the Spotlight was their first major project together.
Cozy Game Pals
Fear the Spotlight didn’t have a particularly buzzy debut, but a few weeks after it hit Steam, Blumhouse Games president Zach Wood and creative lead Louise Blain happened to spot it on Twitter. Castro told Engadget how it went down:
“Zach found it, and he and Louise Blain sat down and played it together and were like, oh my gosh, this is exactly the kind of game that we want to be publishing, this is really great. Let's reach out to them and see, do they need any help? Is there something here that we can work together on?”
“Meanwhile, yes, we did need help,” Singh added, laughing. “We had released it, but we had no idea how to get people to know about it. The people that were finding it were saying very positive things, and we're like, OK, that's great, but now what do we do? We know how to make things, but we don't know anything else about marketing.”
Blumhouse signed Cozy Game Pals and asked how they could help improve Fear the Spotlight. At first, Castro and Singh suggested porting it to consoles and adding additional languages, basic things to get the existing game in front of more players.
“They were excited about the idea, but then they also offered more time,” Singh said. “They asked, what would you do if you had an extra year to work on it?”
The opportunity to expand Fear the Spotlight caught Castro and Singh by surprise. It also scared them, at first.
Cozy Game Pals
“We had never really considered a significant addition to the game before that,” Singh said. “And we also had what we thought was a finished game that we were really proud of. So it was really, really difficult to figure out how to add to a thing that we felt was finished; we didn't want to ruin it. Part of it is our taste and our work, but also part of it feels like black magic. Like, if we mess with it, is it going to come out in a way that we're proud of?”
Castro and Singh took the chance. On October 26, 2023, about one month after Fear the Spotlight’s debut, they removed it from Steam with the promise that they’d add new gameplay, console versions and localization features. They didn’t mention Blumhouse at the time. Behind the scenes, Blumhouse Games gave Cozy Game Pals one year to create the definitive version of Fear the Spotlight, with no creative restrictions.
The revamped version of Fear the Spotlight came out on Steam, PS4, PS5, Switch and Xbox Series X/S on October 22, 2024, developed by Cozy Game Pals and published by Blumhouse Games. It’s the first game in Blumhouse’s publishing roster, which includes future titles from EYES OUT, Half Mermaid, Perfect Garbage, Playmestudio and Vermila Studios.
Cozy Game Pals used the year of extra development time well. Rather than messing with the black magic of the original, Castro and Singh added an entirely new segment, doubling the game’s run time and expanding on their initial ideas in sophisticated, extra-horrific ways. Fear the Spotlight, by the way, is an excellent horror experience. It has low-poly environments, low-res textures and grainy CRT effects, but its animations are smooth and the camera uses friendly third-person controls, nailing the nostalgia without compromising modern conveniences. The story revolves around two teenage friends, Vivian and Amy, and takes them on individual but connected journeys through twisted, spirit-infested versions of reality. Their dialogue and personalities feel authentic, and their emotions are incredibly relatable, whether in the face of unspeakable horrors or just when talking to a crush. It has a few good jump scares, too.
Cozy Game Pals
The first half of Fear the Spotlight is packed with satisfying puzzles, spooky phantoms and tense hide-and-seek mechanics. The second half, created after Blumhouse’s intervention, adds layers of emotional depth and introduces a truly terrifying foe. Vivian is the main playable character in the original version and Amy's story takes center stage in the expanded content.
“The first Vivian story was really us figuring out how to make this game,” Castro said. “But then by the time we were making Amy's, we had so many lessons learned. I feel like the monster is better, the puzzles are better, the storytelling is more streamlined. The second half wraps it up really nicely.”
On top of handling the art, Castro was the main writer on Fear the Spotlight, while Singh handled programming. Castro was the diehard horror fan in the relationship — he was a Resident Evil boy, she was a Silent Hill girl (read to the theme of Avril Lavigne’s Sk8er Boi) — and together, they wanted to capture the fun of being scared in video game form. Fear the Spotlight draws from their personal lives and memories of high school, when every emotion felt new, extreme and sometimes silly. From this lens, Fear the Spotlight also deftly handles serious topics like loss, death, prejudice and love.
“It's just such an impactful time in our, in most people's lives,” Castro said. “I grew up playing these games in the ’90s or in the early 2000s, like Silent Hill one and two and three. So thinking back to high school and how I felt, writing the story was just like, I can only write from my own personal experience. Having a crush and feeling awkward, and when you actually bond with someone, how special that is.”
Singh continued the thread, saying, “I think the home-life stuff — we bond over a lot of our shared experience, which is also represented in the game. Families are complicated, family structures. Having a father that's not present in your life, or the loss of a very close family member, it just changes you and affects you. Those are just pulled from our lives.”
Castro and Singh lovingly described Blumhouse Games as a scrappy team of horror fans, with fewer than 10 people supporting a handful of projects at once, and doing so while trying to prove themselves in a new market. On top of handling trailers and press for Fear the Spotlight’s re-release, the Blumhouse crew helped Cozy Game Pals find a contractor to do a logo and key art, a porting company to help get the game on consoles, and a localization team. More than any of that, though, Castro and Singh said the people at Blumhouse Games seem to truly enjoy the projects they’ve signed.
“They've just been the ideal partner, incredibly supportive,” Castro said. “They really let us decide everything for our game, the game is completely our vision. We would show them prototypes and level designs and of course, they had feedback and thoughts, but yeah — ”
“They know our game really well,” Singh said. “They're genuine fans of the original release. They know our game intimately and can talk to us about our ideas from a very informed perspective.”
Castro concluded, “They come from it from a support perspective. Like, how can we help you guys create your vision that you care about, that you're happy with. It's been amazing.”
Fear the Spotlight is available now for $20 on Steam, PS4, PS5, Switch and Xbox Series X/S.
This article originally appeared on Engadget at https://www.engadget.com/gaming/how-fear-the-spotlight-became-blumhouses-first-video-game-140044877.html?src=rss
In iOS 18.2, Apple is adding a new feature that resurrects some of the intent behind its halted CSAM scanning plans — this time, without breaking end-to-end encryption or providing government backdoors. Rolling out first in Australia, the company’s expansion of its Communication Safety feature uses on-device machine learning to detect and blur nude content, adding warnings and requiring confirmation before users can proceed. If the child is under 13, they can’t continue without entering the device’s Screen Time passcode.
If the device’s onboard machine learning detects nude content, the feature automatically blurs the photo or video, displays a warning that the content may be sensitive and offers ways to get help. The choices include leaving the conversation or group thread, blocking the person and accessing online safety resources.
The feature also displays a message that reassures the child that it’s okay not to view the content or leave the chat. There’s also an option to message a parent or guardian. If the child is 13 or older, they can still confirm they want to continue after receiving those warnings — with a repeat of the reminders that it’s okay to opt out and that further help is available. According toThe Guardian, it also includes an option to report the images and videos to Apple.
Apple
The feature analyzes photos and videos on iPhone and iPad in Messages, AirDrop, Contact Posters (in the Phone or Contacts app) and FaceTime video messages. In addition, it will scan “some third-party apps” if the child selects a photo or video to share with them.
The supported apps vary slightly on other devices. On Mac, it scans messages and some third-party apps if users choose content to share through them. On the Apple Watch, it covers Messages, Contact Posters and FaceTime video messages. Finally, on Vision Pro, the feature scans Messages, AirDrop and some third-party apps (under the same conditions mentioned above).
The feature requires iOS 18, iPadOS 18, macOS Sequoia or visionOS 2.
The Guardian reports that Apple plans to expand it globally after the Australia trial. The company likely chose the land Down Under for a specific reason: The country is set to roll out new regulations that require Big Tech to police child abuse and terror content. As part of the new rules, Australia agreed to add the clause that it was only mandated “where technically feasible,” omitting a requirement to break end-to-end encryption and compromise security. Companies will need to comply by the end of the year.
User privacy and security were at the heart of the controversy over Apple’s infamous attempt to police CSAM. In 2021, it prepared to adopt a system that would scan for images of online sexual abuse, which would then be sent to human reviewers. (It came as something of a shock after Apple’s history of standing up to the FBI over its attempts to unlock an iPhone belonging to a terrorist.) Privacy and security experts argued that the feature would open a backdoor for authoritarian regimes to spy on their citizens in situations without any exploitative material. The following year, Apple abandoned the feature, leading (indirectly) to the more balanced child-safety feature announced today.
Once it rolls out globally, you can activate the feature under Settings > Screen Time > Communication Safety, and toggle the option on. That section has been activated by default since iOS 17.
This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/ios-182-has-a-child-safety-feature-that-can-blur-nude-content-and-report-it-to-apple-194614810.html?src=rss
If you've been looking to upgrade your smartwatch, now might be the right time. Our favorite smartwatch, the Apple Watch Series 10, is currently on sale with the 42mm model down to $370 from $399 in Jet Black. The 46mm model is down to $399 from $429 and available in Jet Black and Silver Aluminum.
Apple released its Series 10 Watch only a month ago, alongside new products like the AirPods 4 and the iPhone 16. We gave it a 90 in our review features like its thinner frame and bigger screens — from 41mm to 42mm and 45mm to 46mm. The watch also offers music and podcast listening through its onboard speaker, which worked alright for a quick listen in a quiet space. Its battery also lasts slightly longer than its predecessor's.
Health-wise, the Apple Watch Series 10 brings sleep apnea monitoring, tracking your sleeping patterns for 30 days to see if its worth going to a doctor. Notably, the Series 10 Watch doesn't include a Blood Oxygen app as Apple faces ongoing copyright litigation arguing the tech giant infringed on a patent owned by health tech company, Massimo.
Follow @EngadgetDeals on Twitter and subscribe to the Engadget Deals newsletter for the latest tech deals and buying advice.
This article originally appeared on Engadget at https://www.engadget.com/deals/the-apple-watch-series-10-is-30-off-right-now-133026681.html?src=rss
Google just announced a spate of safety features coming to Messages. There’s enhanced scam detection centered around texts that could lead to fraud. The company says the update provides “improved analysis of scammy texts.” For now, this tool will prioritize scams involving package deliveries and job offers.
When Google Messages suspects a scam, it’ll move the message to the spam folder or issue a warning. The app uses on-device machine learning models to detect these scams, meaning that conversations will remain private. This enhancement is rolling out now to beta users who have spam protection enabled.
Google’s also set to broadly roll out intelligent warnings, a feature that’s been in the pilot stage for a while. This tool warns users when they get a link from an unknown sender and automatically “blocks messages with links from suspicious senders.” The updated safety tools also include new sensitive content warnings that automatically blurs images that may contain nudity. This is an opt-in feature and also keeps everything on the device. It’ll show up in the next few months.
Finally, there’s a forthcoming tool that’ll let people turn off messages from unknown international senders, thus cutting the scam spigot off at the source. This will automatically hide messages from international senders who aren’t already in the contacts list. This feature is entering a pilot program in Singapore later this year before expanding to more countries.
In addition to the above tools, Google says it’s currently working on a contact verifying feature for Android. This should help put the kibosh on scammers trying to impersonate one of your contacts. The company has stated that this feature will be available sometime next year.
This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/google-messages-adds-enhanced-scam-detection-tools-190009890.html?src=rss