Google is counting on its very own GPT-4 competitor, Gemini, so much that it staged parts of a recent demo video. In an opinion piece, Bloomberg says Google admits that for its video titled "Hands-on with Gemini: Interacting with multimodal AI," not only was it edited to speed up the outputs (which was declared in the video description), but the implied voice interaction between the human user and the AI was actually non-existent. 

Instead, the actual demo was made by "using still image frames from the footage, and prompting via text," rather than having Gemini respond to — or even predict — a drawing or change of objects on the table in real time. This is far less impressive than the video wants to mislead us into thinking, and worse yet, the lack of disclaimer about the actual input method makes Gemini's readiness rather questionable.

It comes as no surprise that Google denies any wrongdoing here, as it referred The Verge to an X post written by Gemini's co-lead, Oriol Vinyals, which says "all the user prompts and outputs in the video are real," and that his team made the video "to inspire developers." Given the industry and authorities' attention on AI lately, perhaps the tech giant should be more sensitive about its presentations in this field.

ByteDance's TikTok has been facing a tough time over its handling of Israel-Hamas war content and related hate speech, but it is now stepping up with some new initiatives. On the moderation side, the social platform is rolling out new comment filtering tools, with the most notable one being "Comment Care Mode," which supposedly automatically filters comments that are deemed similar to the creator's previously reported or deleted ones. Another new feature helps nix comments made by accounts that are not in the creator's following or follower list. The company aims to increase new users' awareness of these tools via a prompt after their first video upload, and in the long run, it'll set up a product beta testing program to get direct feedback from creators.

TikTok has also set up a new anti-hate and discrimination task force, in the hopes of proactively spotting antisemitism, Islamophobia and other hate trends before they get out of hand. The team will work with experts on improving training for moderators to better address hate speech, and it will expand its managed creator communities to Jewish plus other inter-faith communities, as well as API and LGBTQ+, next year. 

The Information added that TikTok plans to expand access to its research APIs to civil society groups — as the likes of the Anti-Defamation League have been requesting for years, apparently — so they can better understand the types of content spreading on TikTok. This comes in stark contrast to how X — well, Elon Musk, mostly — limited social media researchers' access to its platform, while it continues to deny any wrongdoing over accusations of antisemetic content.

While TikTok's stepped-up efforts may not convince those who still accuse its algorithm of bias, the platform has at least continued removing a staggering amount of offending content. The latest figure on removed videos in the conflict region has hit 1.3 million, between October 7 and November 30. These included "content promoting Hamas, hate speech, terrorism and misinformation."

Today, Meta has unveiled what it calls "the biggest set of improvements to Messenger since it was first launched in 2011." First and foremost, end-to-end encryption is now a default for private chats and calls on Messenger and Facebook, meaning your secured communication can't be spied on by others — not even by Meta itself, apparently. Once updated, Messenger will ask users to set up a PIN, in case they need to recover messages on a new device later. The global rollout may take a few months to complete, due to the fact that the app has over a billion users.

End-to-end encryption became an option for Messenger in 2016, but Meta is obviously stepping up its safety efforts — a sensible move given the company's other recent scandals, especially with child predation content. Meta added that "this has taken years to deliver because we’ve taken our time to get this right," and also "to rebuild Messenger features from the ground up."

Meta

In addition to a set of new privacy, safety and control features, Messenger is getting improved image quality for photos and videos. Meta says it's currently testing HD media sharing with a small test group, before rolling this out "in the coming months." Messenger is also receiving other handy tools that are seemingly inspired by WhatsApp, namely message edit (for up to 15 minutes after sending), voice message playback speed options, continued voice message playback outside the chat or app, read receipts control and disappearing messages (after 24 hours; this is now available to all chats since end-to-end encryption has become a default).

Even though Windows 10 is still getting the Copilot AI love, the fact remains that it will no longer receive updates as of its end-of-support date, October 14, 2025, as noted in Microsoft's IT Pro Blog post. Still, to ensure companies — and potentially individual consumers later — have ample time to prepare for the eventual upgrade, Microsoft will soon offer an Extended Security Update (ESU) program for Windows 10.

Much like the similar program made for Windows 7, organizations can buy Extended Security Updates for Windows 10 — now locked at version 22H2 — by way of a yearly subscription, with the maximum extended life being three years, i.e. up to October 14, 2028. Note that the program only covers critical and important security updates, and that there will be no technical support beyond these patches.

An alternative solution to the above is to migrate the Windows 10 PCs to Windows 11 in the cloud, by way of Windows 365 subscription. That way, the actual Windows 10 system in these machines will benefit from the Extended Security Updates at no extra cost, but still only for three years maximum.

Of course, Microsoft would much rather you just upgrade to Windows 11 one way or another, but as pointed out by Ars Technica, this time the tech giant plans on extending the ESU program to individuals. This wasn't the case with Windows 7's ESU program, which goes to show that Microsoft is well aware of Windows 10's dominance even today — as much as 68.02 percent, versus just 26.63 percent for Windows 11, according to Statcounter (as of November 2023). Stay tuned for further details and pricing later.

Back in August, Meta unveiled its multimodal AI translation model, SeamlessM4T, which supports nearly 100 languages for text and 36 for speech. With an updated "v2" architecture, the tech giant is now expanding on this tool to make conversational translations more spontaneous and expressive — the latter a missing key to an authentic conversation across languages.

The first of the two new features is "SeamlessExpressive" which, as you can tell by the name, ports your expressions over to your translated speech. These include your pitch, volume, emotional tone (excitement, sadness or whispers), speech rate and pauses. Considering how translated speeches had always sounded robotic until now, this breakthrough is potentially a game-changer — both in our daily lives and also in content production. Supported languages include English, Spanish, German, French, Italian and Chinese, though the demo page is missing Italian and Chinese at the time of writing this article.

The second feature is "SeamlessStreaming," which starts translating a speech while the speaker is still talking, thus allowing others to hear a translation faster. There's still a short latency of just under two seconds, but at least you won't have to wait until someone finishes a sentence. According to Meta, the challenge here is that different languages have different sentence structures, so it had to develop an algorithm dedicated to studying partial audio input, in order to decide whether there's enough context to start generating a translated output, or whether to keep listening.

Meta's latest development on this "Seamless Communication" suite seems to be an impressive one — more so than the mobile interpreter tools offered by the likes of Google and Samsung. There's no word on when the public will be able to utilize these new features, but I can already imagine Meta baking them into its smart glasses some day, making them even more practical than ever.