A U.S. federal judge has decided that free-speech protections in the First Amendment "don't shield an AI company from a lawsuit," reports Legal Newsline.

The suit is against Character.AI (a company reportedly valued at $1 billion with 20 million users) Judge Anne C. Conway of the Middle District of Florida denied several motions by defendants Character Technologies and founders Daniel De Freitas and Noam Shazeer to dismiss the lawsuit brought by the mother of 14-year-old Sewell Setzer III. Setzer killed himself with a gun in February of last year after interacting for months with Character.AI chatbots imitating fictitious characters from the Game of Thrones franchise, according to the lawsuit filed by Sewell's mother, Megan Garcia.

"... Defendants fail to articulate why words strung together by (Large Language Models, or LLMs, trained in engaging in open dialog with online users) are speech," Conway said in her May 21 opinion. "... The court is not prepared to hold that Character.AI's output is speech."
Character.AI's spokesperson told Legal Newsline they've now launched safety features (including an under-18 LLM, filter Characters, time-spent notifications and "updated prominent disclaimers" (as well as a "parental insights" feature). "The company also said it has put in place protections to detect and prevent dialog about self-harm. That may include a pop-up message directing users to the National Suicide and Crisis Lifeline, according to Character.AI."

Thanks to long-time Slashdot reader schwit1 for sharing the news.

"We've officially entered the age of watching robots clobber each other in fighting rings," writes Vice.com.

A kick-boxing competition was staged Sunday in Hangzhou, China using four robots from Unitree Robotics, reports Futurism. (The robots were named "AI Strategist", "Silk Artisan", "Armored Mulan", and "Energy Guardian".) "However, the robots weren't acting autonomously just yet, as they were being remotely controlled by human operator teams."

Although those ringside human controllers used quick voice commands, according to the South China Morning Post: Unlike typical remote-controlled toys, handling Unitree's G1 robots entails "a whole set of motion-control algorithms powered by large [artificial intelligence] models", said Liu Tai, deputy chief engineer at China Telecommunication Technology Labs, which is under research institute China Academy of Information and Communications Technology.
More from Vice: The G1 robots are just over 4 feet tall [130 cm] and weigh around 77 pounds [35 kg]. They wear gloves. They have headgear. They throw jabs, uppercuts, and surprisingly sharp kicks... One match even ended in a proper knockout when a robot stayed down for more than eight seconds. The fights ran three rounds and were scored based on clean hits to the head and torso, just like standard kickboxing...
Thanks to long-time Slashdot reader AmiMoJo for sharing the news.

Thursday Anthropic's CEO/cofounder Dario Amodei again warned unemployment could spike 10 to 20% within the next five years as AI potentially eliminated half of all entry-level white-collar jobs.

But CNN's senior business writer dismisses that as "all part of the AI hype machine," pointing out that Amodei "didn't cite any research or evidence for that 50% estimate." And that was just one of many of the wild claims he made that are increasingly part of a Silicon Valley script: AI will fix everything, but first it has to ruin everything. Why? Just trust us.

In this as-yet fictional world, "cancer is cured, the economy grows at 10% a year, the budget is balanced — and 20% of people don't have jobs," Amodei told Axios, repeating one of the industry's favorite unfalsifiable claims about a disease-free utopia on the horizon, courtesy of AI. But how will the US economy, in particular, grow so robustly when the jobless masses can't afford to buy anything? Amodei didn't say... Anyway. The point is, Amodei is a salesman, and it's in his interest to make his product appear inevitable and so powerful it's scary. Axios framed Amodei's economic prediction as a "white-collar bloodbath."

Even some AI optimists were put off by Amodei's stark characterization. "Someone needs to remind the CEO that at one point there were more than (2 million) secretaries. There were also separate employees to do in office dictation," wrote tech entrepreneur Mark Cuban on Bluesky. "They were the original white collar displacements. New companies with new jobs will come from AI and increase TOTAL employment."

Little of what Amodei told Axios was new, but it was calibrated to sound just outrageous enough to draw attention to Anthropic's work, days after it released a major model update to its Claude chatbot, one of the top rivals to OpenAI's ChatGPT.
Amodei told CNN Thursday this great societal change would be driven by how incredibly fast AI technology is getting better and better — and that the AI boom "is bigger and it's broader and it's moving faster than anything has before...!"

An anonymous reader quotes a report from SiliconANGLE: Hugging Face has open-sourced the blueprints of two internally developed robots called HopeJR and Reachy Mini. The company debuted the machines on Thursday. Hugging Face is backed by more than $390 million in funding from Nvidia Corp., IBM Corp. and other investors. It operates a GitHub-like platform for sharing open-source artificial intelligence projects. It says its platform hosts more than 1 million AI models, hundreds of thousands of datasets and various other technical assets.

The company started prioritizing robotics last year after launching LeRobot, a section of its platform dedicated to autonomous machines. The portal provides access to AI models for powering robots and datasets that can be used to train those models. Hugging Face released its first hardware blueprint, a robotic arm design called the SO-100, late last year. The SO-100 was developed in partnership with a startup called The Robot Studio. Hugging Face also collaborated with the company on the HopeJR, the first new robot that debuted this week. According to TechCrunch, it's a humanoid robot that can perform 66 movements including walking.

HopeJR is equipped with a pair of robotic arms that can be remotely controlled by a human using a pair of specialized, chip-equipped gloves. HopeJR's arms replicate the movements made by the wearer of the gloves. A demo video shared by Hugging Face showed that the robot can shake hands, point to a specific text snippet on a piece of paper and perform other tasks. Hugging Face's other new robot, the Reachy Mini, likewise features an open-source design. It's based on technology that the company obtained through the acquisition of a venture-backed startup called Pollen Robotics earlier this year. Reachy Mini is a turtle-like robot that comes in a rectangular case. Its main mechanical feature is a retractable neck that allows it to follow the user with its head or withdraw into the case. This case, which is stationary, is compact and lightweight enough to be placed on a desk. Hugging Face will offer pre-assembled versions of its open-source Reach Mini and HopeJR robots for $250 and $3,000, with the first units starting to ship by the end of the year.

Artificial intelligence could soon outpace Bitcoin mining in energy consumption, according to Alex de Vries-Gao, a PhD candidate at Vrije Universiteit Amsterdam's Institute for Environmental Studies. His research estimates that by the end of 2025, AI could account for nearly half of all electricity used by data centers worldwide -- raising significant concerns about its impact on global climate goals.

"While companies like Google and Microsoft disclose total emissions, few provide transparency on how much of that is driven specifically by AI," notes DIGIT. To fill this gap, de Vries-Gao employed a triangulation method combining chip production data, corporate disclosures, and industry analyst estimates to map AI's growing energy footprint.

His analysis suggests that specialized AI hardware could consume between 46 and 82 terawatt-hours (TWh) in 2025 -- comparable to the annual energy usage of countries like Switzerland. Drawing on supply chain data, the study estimates that millions of AI accelerators from NVIDIA and AMD were produced between 2023 and 2024, with a potential combined power demand exceeding 12 gigawatts (GW). A detailed explanation of his methodology is available in his commentary published in Joule.

In a full-circle moment for Palmer Luckey, Meta and his defense tech company Anduril are teaming up to develop mixed reality headsets for the U.S. military under the Army's revamped SBMC Next program. The collaboration will merge Meta's Reality Labs hardware and Llama AI with Anduril's battlefield software, marking Meta's entry into military XR through the very company founded by Luckey after his controversial departure from Facebook. "I am glad to be working with Meta once again," Luckey said in a blog post. "My mission has long been to turn warfighters into technomancers, and the products we are building with Meta do just that." TechCrunch reports: This partnership stems from the Soldier Borne Mission Command (SBMC) Next program, formerly called the Integrated Visual Augmentation System (IVAS) Next. IVAS was a massive military contract, with a total $22 billion budget, originally awarded to Microsoft in 2018 intended to develop HoloLens-like AR glasses for soldiers. But after endless problems, in February the Army stripped management of the program from Microsoft and awarded it to Anduril, with Microsoft staying on as a cloud provider. The intent is to eventually have multiple suppliers of mixed reality glasses for soldiers.

All of this meant that if Luckey's former employer, Meta, wanted to tap into the potentially lucrative world of military VR/AR/XR headsets, it would need to go through Anduril. The devices will be based on tech out of Meta's AR/VR research center Reality Labs, the post says. They'll use Meta's Llama AI model, and they will tap into Anduril's command and control software known as Lattice. The idea is to provide soldiers with a heads-up display of battlefield intelligence in real time. [...] An Anduril spokesperson tells TechCrunch that the product family Meta and Anduril are building is even called EagleEye, which will be an ecosystem of devices. EagleEye is what Luckey named Anduril's first imagined headset in Anduril's pitch deck draft, before his investors convinced him to focus on building software first. After the announcement, Luckey said on X: "It is pretty cool to have everything at our fingertips for this joint effort -- everything I made before Meta acquired Oculus, everything we made together, and everything we did on our own after I was fired."

An anonymous reader quotes a report from 404 Media: If you've left a comment on a YouTube video, a new website claims it might be able to find every comment you've ever left on any video you've ever watched. Then an AI can build a profile of the commenter and guess where you live, what languages you speak, and what your politics might be. The service is called YouTube-Tools and is just the latest in a suite of web-based tools that started life as a site to investigate League of Legends usernames. Now it uses a modified large language model created by the company Mistral to generate a background report on YouTube commenters based on their conversations. Its developer claims it's meant to be used by the cops, but anyone can sign up. It costs about $20 a month to use and all you need to get started is a credit card and an email address.

The tool presents a significant privacy risk, and shows that people may not be as anonymous in the YouTube comments sections as they may think. The site's report is ready in seconds and provides enough data for an AI to flag identifying details about a commenter. The tool could be a boon for harassers attempting to build profiles of their targets, and 404 Media has seen evidence that harassment-focused communities have used the developers' other tools. YouTube-Tools also appears to be a violation of YouTube's privacy policies, and raises questions about what YouTube is doing to stop the scraping and repurposing of peoples' data like this. "Public search engines may scrape data only in accordance with YouTube's robots.txt file or with YouTube's prior written permission," it says.

Lovable, a Swedish startup that allows users to create websites and apps through natural language prompts, failed to address a critical security vulnerability for months after being notified, according to a new report. A study by Replit employees found that 170 of 1,645 Lovable-created applications exposed sensitive user information including names, email addresses, financial data, and API keys that could allow hackers to run up charges on customers' accounts.

The vulnerability, published this week in the National Vulnerabilities Database, stems from misconfigured Supabase databases that Lovable's AI-generated code connects to for storing user data. Despite being alerted to the problem in March, Lovable initially dismissed concerns and only later implemented a limited security scan that checks whether database access controls are enabled but cannot determine if they are properly configured.

An anonymous reader shares a report: Some of the citations that underpin the science in the White House's sweeping "MAHA Report" appear to have been generated using artificial intelligence [non-paywalled source], resulting in numerous garbled scientific references and invented studies, AI experts said Thursday. Of the 522 footnotes to scientific research in an initial version of the report sent to The Washington Post, at least 37 appear multiple times, according to a review of the report by The Post.

Other citations include the wrong author, and several studies cited by the extensive health report do not exist at all, a fact first reported by the online news outlet NOTUS on Thursday morning. Some references include "oaicite" attached to URLs -- a definitive sign that the research was collected using artificial intelligence. The presence of "oaicite" is a marker indicating use of OpenAI, a U.S. artificial intelligence company. A common hallmark of AI chatbots, such as ChatGPT, is unusually repetitive content that does not sound human or is inaccurate -- as well as the tendency to "hallucinate" studies or answers that appear to make sense but are not real.

Google has begun automatically generating AI-powered email summaries for Gmail Workspace users, eliminating the need to manually trigger the feature that has been available since last year. The company's Gemini AI will now independently determine when longer email threads or messages with multiple replies would benefit from summarization, displaying these summaries above the email content itself. The automatic summaries currently appear only on mobile devices for English-language emails and may take up to two weeks to roll out to individual accounts, with Google providing no timeline for desktop expansion or availability to non-Workspace Gmail users.

An anonymous reader shares a report: Negotiations are under way between the UAE and OpenAI that may make the company's ChatGPT Plus artificial intelligence chatbot available to all residents free of charge, though a final deal has not been reached.

An agreement involving ChatGPT Plus would be part of the recently announced Stargate UAE infrastructure plan to create an AI hub in Abu Dhabi, according to a source familiar with the country's AI strategy. Abu Dhabi's AI company G42 has partnered with OpenAI, Oracle and Nvidia to set up Stargate UAE, a 1-gigawatt computing cluster that will operate in the newly established 5GW UAE -- US AI Campus.

An anonymous reader quotes a report from The Guardian: Whether it is doing sums or working out what to text your new date, some tasks produce a furrowed brow. Now scientists say they have come up with a device to monitor such effort: an electronic tattoo, stuck to the forehead. The researchers say the device could prove valuable among pilots, healthcare workers and other professions where managing mental workload is crucial to preventing catastrophes. "For this kind of high-demand and high-stake scenario, eventually we hope to have this real-time mental workload decoder that can give people some warning and alert so that they can self-adjust, or they can ask AI or a co-worker to offload some of their work," said Dr Nanshu Lu, an author of the research from the University of Texas at Austin, adding the device may not only help workers avoid serious mistakes but also protect their health.

Writing in the journal Device, Lu and colleagues describe how using questionnaires to investigate mental workload is problematic, not least as people are poor at objectively judging cognitive effort and they are usually conducted after a task. Meanwhile, existing electroencephalography (EEG) and electrooculography (EOG) devices, that can be used to assess mental workload by measuring brain waves and eye movements respectively, are wired, bulky and prone to erroneous measurements arising from movements. By contrast, the "e-tattoo" is a lightweight, flexible, wireless device.

The black, wiggly path of the e-tattoo is composed of a graphite-based conductive material, and is attached to the forehead using conductive adhesive film. Four square EEG electrodes, positioned on the forehead, each detect a different region of brain activity -- with a reference electrode behind the ear -- while rectangular EOG electrodes, placed vertically and horizontally around the eyes, provide data about eye movements. Each of the stretchable electrodes is coated in an additional conductive material. The e-tattoo, which is bespoke and disposable, is connected to a reusable flexible printed circuit using conductive tape, while a lightweight battery can be clipped to the device. The device is expected to cost less than $200 and be accompanied with an app to alert the user if their mental workload is too high.

Google's Gemini AI can now analyze and summarize video files stored in Google Drive, letting users ask questions about content like meeting takeaways or product updates without watching the footage. The Verge reports: The Gemini in Drive feature provides a familiar chatbot interface that can provide quick summaries describing the footage or pull specific information. For example, users can ask Gemini to list action items mentioned in recorded meetings or highlight the biggest updates and new products in an announcement video, saving time spent on manually combing through and taking notes.

The feature requires captions to be enabled for videos, and can be accessed using either Google Drive's overlay previewer or a new browser tab window. It's available in English for Google Workspace and Google One AI Premium users, and anyone who has previously purchased Gemini Business or Enterprise add-ons, though it may take a few weeks to fully roll out. You can learn more about the update in Google's blog post.

An anonymous reader quotes a report from SC Media: Thousands of ASUS routers have been compromised with malware-free backdoors in an ongoing campaign to potentially build a future botnet, GreyNoise reported Wednesday. The threat actors abuse security vulnerabilities and legitimate router features to establish persistent access without the use of malware, and these backdoors survive both reboots and firmware updates, making them difficult to remove.

The attacks, which researchers suspect are conducted by highly sophisticated threat actors, were first detected by GreyNoise's AI-powered Sift tool in mid-March and disclosed Thursday after coordination with government officials and industry partners. Sekoia.io also reported the compromise of thousands of ASUS routers in their investigation of a broader campaign, dubbed ViciousTrap, in which edge devices from other brands were also compromised to create a honeypot network. Sekoia.io found that the ASUS routers were not used to create honeypots, and that the threat actors gained SSH access using the same port, TCP/53282, identified by GreyNoise in their report. The backdoor campaign affects multiple ASUS router models, including the RT-AC3200, RT-AC3100, GT-AC2900, and Lyra Mini.

GreyNoise advises users to perform a full factory reset and manually reconfigure any potentially compromised device. To identify a breach, users should check for SSH access on TCP port 53282 and inspect the authorized_keys file for unauthorized entries.

DevNull127 writes: Stack Overflow will test paying experts to answer questions. That's one of many radical experiments they're now trying to stave off an AI-induced death spiral. Questions and answers to the site have plummeted more than 90% since April of 2020. So here's what Stack Overflow will try next.

1. They're bringing back Chat, according to their CEO (to foster "even more connections between our community members" in "an increasingly AI-driven world").

2. They're building a "new Stack Overflow" meant to feel like a personalized portal. "It might collect videos, blogs, Q&A, war stories, jokes, educational materials, jobs... and fold them together into one personalized destination."

3. They're proposing areas more open to discussion, described as "more flexible Stack Exchanges... where users can explore ideas or share opinions."

4. They're also licensing Stack Overflow content to AI companies for training their models.

5. Again, they will test paying experts to answer questions.

Anthropic co-founder and CEO Dario Amodei is warning that AI could eliminate half of all entry-level white-collar jobs within the next five years -- and overall unemployment potentially spiking between 10 and 20% during that period.

The prediction comes as new data from venture capital firm SignalFire shows Big Tech companies have already reduced their hiring of new graduates by approximately 50% compared to pre-pandemic levels, with AI adoption cited as a contributing factor. Amodei told Axios that AI companies and government officials are "sugarcoating" the risks of mass job displacement in technology, finance, law, and consulting sectors.

Arizona State University researchers are pushing back [PDF] against the widespread practice of describing AI language models' intermediate text generation as "reasoning" or "thinking," arguing this anthropomorphization creates dangerous misconceptions about how these systems actually work. The research team, led by Subbarao Kambhampati, examined recent "reasoning" models like DeepSeek's R1, which generate lengthy intermediate token sequences before providing final answers to complex problems. Though these models show improved performance and their intermediate outputs often resemble human scratch work, the researchers found little evidence that these tokens represent genuine reasoning processes.

Crucially, the analysis also revealed that models trained on incorrect or semantically meaningless intermediate traces can still maintain or even improve performance compared to those trained on correct reasoning steps. The researchers tested this by training models on deliberately corrupted algorithmic traces and found sustained improvements despite the semantic noise. The paper warns that treating these intermediate outputs as interpretable reasoning traces engenders false confidence in AI capabilities and may mislead both researchers and users about the systems' actual problem-solving mechanisms.

An anonymous reader quotes a report from TechCrunch: Researchers at SignalFire, a data-driven VC firm that tracks job movements of over 600 million employees and 80 million companies on LinkedIn, believe they may be seeing first signs of AI's impact on hiring. When analyzing hiring trends, SignalFire noticed that tech companies recruited fewer recent college graduates in 2024 than they did in 2023. Meanwhile, tech companies, especially the top 15 Big Tech businesses, ramped up their hiring of experienced professionals. Specifically, SignalFire found that Big Tech companies reduced the hiring of new graduates by 25% in 2024 compared to 2023. Meanwhile, graduate recruitment at startups decreased by 11% compared to the prior year. Although SignalFire wouldn't reveal exactly how many fewer grads were hired according to their data, a spokesperson told us it was thousands.

While adoption of new AI tools might not fully explain the dip in recent grad hiring, Asher Bantock, SignalFire's head of research, says there's "convincing evidence" that AI is a significant contributing factor. Entry-level jobs are susceptible to automation because they often involve routine, low-risk tasks that generative AI handles well. AI's new coding, debugging, financial research, and software installation abilities could mean companies need fewer people to do that type of work. AI's ability to handle certain entry-level tasks means some jobs for new graduates could soon be obsolete. [...]

Although AI's threat to low-skilled jobs is real, tech companies' need for experienced professionals is still rising. According to SignalFire's report, Big Tech companies increased hiring by 27% for professionals with two to five years of experience, while startups hired 14% more individuals in that same seniority range. A frustrating paradox emerges for recent graduates: They can't get hired without experience, but they can't get experience without being hired. While this dilemma is not new, Heather Doshay, SignalFire's people and talent partner, says it is considerably exacerbated by AI. Doshay's advice to new grads: master AI tools. "AI won't take your job if you're the one who's best at using it," she said.

An anonymous reader quotes a report from The Verge: Secretary of State Marco Rubio announced Wednesday that the U.S. would restrict visas for "foreign nationals who are responsible for censorship of protected expression in the United States." He called it "unacceptable for foreign officials to issue or threaten arrest warrants on U.S. citizens or U.S. residents for social media posts on American platforms while physically present on U.S. soil" and "for foreign officials to demand that American tech platforms adopt global content moderation policies or engage in censorship activity that reaches beyond their authority and into the United States."

It's not yet clear how or against whom the policy will be enforced, but seems to implicate Europe's Digital Services Act, a law that came into effect in 2023 with the goal of making online platforms safer by imposing requirements on the largest platforms around removing illegal content and providing transparency about their content moderation. Though it's not mentioned directly in the press release about the visa restrictions, the Trump administration has slammed the law on multiple occasions, including in remarks earlier this year by Vice President JD Vance.

The State Department's homepage currently links to an article on its official Substack, where senior advisor for the Bureau of Democracy, Human Rights, and Labor Samuel Samson critiques the DSA as a tool to "silence dissident voices through Orwellian content moderation." He adds, "Independent regulators now police social media companies, including prominent American platforms like X, and threaten immense fines for non-compliance with their strict speech regulations." "We will not tolerate encroachments upon American sovereignty," Rubio says in the announcement, "especially when such encroachments undermine the exercise of our fundamental right to free speech."

Telegram has partnered with xAI to integrate the Grok chatbot into its platform for one year, with xAI paying $300 million in cash and equity. Telegram will also receive 50% of subscription revenue from Grok. TechCrunch reports: Earlier this year, xAI made the Grok chatbot available to Telegram's premium users. It seems Grok might now be made available to all users. A video posted by [Telegram CEO Pavel Durov] on X suggested that Grok can be pinned on top of chats within the app, and users can also ask questions to Grok from the search bar. Notably, Meta has also integrated Meta AI into the search bar on Instagram and WhatsApp. The video also shows that you will be able to use Grok for writing suggestions, summarizing chats, links, and documents, and creating stickers. Grok will supposedly also help answer questions for businesses and assist with moderation. UPDATE: In a response to Durov's X post outlining the partnership, Elon Musk said: "No deal has been signed."

"Musk's denial, however, raises questions about the status and structure of the agreement," reports TheStreet. "It's unclear whether the partnership has been formalized or if Durov was announcing a framework that remains under discussion. Neither Telegram nor xAI has issued a follow-up clarification."