Hackers used Meta’s AI-powered support chatbot to infiltrate high-profile [Instagram](https://www.theguardian.com/technology/instagram) accounts, the company confirmed on Monday, saying it had resolved the problem after researchers exposed it. The targets ranged from Barack Obama’s White House account to Sephora and the US Space Force Chief Master Sergeant, according to reporting from [404 Media](https://www.404media.co/hackers-simply-asked-meta-ai-to-give-them-access-to-high-profile-instagram-accounts-it-worked/). Everyday users complained of similar hijackings [on Reddit and X](https://techcrunch.com/2026/06/01/hackers-hijacked-instagram-accounts-by-tricking-meta-ai-support-chatbot-into-granting-access/) over the weekend. Security researchers and hacking groups posted videos and screenshots of how to steal an account on Telegram, and a video shared on X appears to show a hacker telling Meta’s AI assistant to link the account to a new email address; the bot assures the hacker a verification code has been sent to that new email, and asks to input the numbers in the chat interface. Once the hacker pastes the correct number, they are shown a button to reset the targeted accounts’ password. In at least one video, the hacker used a virtual private network to spoof the account holder’s location and avoid Meta’s safeguards. Meta said in a statement: “This issue has been resolved, and we are securing impacted accounts.” It is unclear how many accounts were affected. The breach raises concerns about just how safe it is to rely on AI for key security measures such as passwords. Meta, which is [rapidly reorganizing](https://www.theguardian.com/technology/2026/may/19/meta-jobs-ai-transfers) its workers’ jobs around AI and increasing the use of AI features in its platforms, [rolled out](https://about.fb.com/news/2026/03/boosting-your-support-and-safety-on-metas-apps-with-ai/) the AI support assistant globally on Facebook and Instagram earlier this year. The press release for the new AI feature explained that the new AI support assistant can “take action for you on a growing set of requests directly within Facebook and in the future, on Instagram”. The list of actions included reporting scams, impersonation accounts, or problematic content, as well as resetting passwords. “The Meta AI support assistant is a major step in our work to deliver stronger support on our apps,” reads a March press release from Meta.

183602006 story [](//meta.slashdot.org/index2.pl?fhfilter=ai)[ ](//meta.slashdot.org/index2.pl?fhfilter=facebook)[](//meta.slashdot.org/index2.pl?fhfilter=meta) Posted by [BeauHD](https://www.linkedin.com/in/beauhd/) on Thursday June 04, 2026 @04:00PM from the losing-ground dept. Meta has reportedly [delayed the developer release of its Muse Spark AI model API multiple times](https://www.reuters.com/technology/meta-repeatedly-pushes-back-new-ai-model-release-developers-wsj-says-2026-06-04/), and as of Tuesday, had no scheduled launch date, according to the [Wall Street Journal](https://www.wsj.com/tech/ai/meta-keeps-delaying-the-release-of-its-new-ai-model-to-developers-f8569c8c?mod=hp_lead_pos3) _(paywalled)._ Reuters reports: _A Meta spokesperson told Reuters on Wednesday that the company is already testing the Application Programming Interface (API) with some early partners and is looking forward to releasing it this month. "The muse spark API will be coming soon," Meta AI Chief Alexandr Wang announced in a post on X in April. Meta unveiled Muse Spark in April as the first model built to close the gap with rivals. Muse Spark is the first in a new series of models created by the company's Superintelligence Labs. Earlier on Wednesday, Meta [unveiled](https://about.fb.com/news/2026/06/meta-business-agent/) an AI agent aimed at helping businesses carry out day-to-day operations, hinting at the company's ambitions to compete with rivals such as OpenAI, Anthropic and Alphabet's Google. _

183871918 story [](//tech.slashdot.org/index2.pl?fhfilter=ai)[](//tech.slashdot.org/index2.pl?fhfilter=facebook) Posted by EditorDavid on Sunday June 14, 2026 @02:43PM from the friend-request dept. "A year after spending over $14 billion to bring in Alexandr Wang and a group of his top Scale AI engineers to revamp its artificial intelligence efforts, Meta is at least back on the map in AI," [reports CNBC](https://www.cnbc.com/2026/06/14/meta-hired-alexandr-wang-to-build-ai-its-zuckerbergs-job-to-sell-it.html), "though it's still far behind OpenAI, Anthropic and Google in the market." _Wang's big accomplishment was the delivery of [the Muse Spark AI model](https://www.cnbc.com/2026/04/08/meta-debuts-first-major-ai-model-since-14-billion-deal-to-bring-in-alexandr-wang.html) in April, marking Meta's first jump into proprietary foundation models and away from a strict adherence to open source, or open weight as it's more commonly called in AI... "Meta needs to provide more proof points of both adoption and commercialization," said Ralph Schackart, an analyst at William Blair who recommends buying the stock. "Investors are looking for Meta to monetize a new AI-first product, beyond the substantial positive impact AI is having on enhancing the advertising models." Wall Street, at least so far, is unimpressed. Meta's stock is down 18% over the past 12 months, the worst performer in the megacap group, along with Microsoft, which has its own challenges in AI. That's even after Meta reported 33% revenue growth in the first quarter, the [fastest rate](https://www.cnbc.com/2026/04/29/meta-q1-earnings-report-2026.html) of expansion for any period since 2021. For Meta, the problem started with what some industry experts called, in hindsight at least, a strategic blunder. The company jumped into AI with its Llama family of models, offering an [open-source approach](https://www.cnbc.com/2023/10/16/metas-open-source-approach-to-ai-puzzles-wall-street-techies-love-it.html) that allowed developers to freely tinker, while the other big model makers charged for access. In April of last year, Meta's release of Llama 4 fell flat, failing to captivate developers and leading Zuckerberg to reconsider his company's approach to AI development... Since the release of Muse Spark, Meta has unveiled new AI and business-related [subscription plans](https://www.cnbc.com/2026/05/27/meta-testing-ai-subscription-services-cheapest-plan-at-7point99-a-month.html) as part of an effort to expand its business beyond online ads. Historically, it [hasn't worked](https://www.cnbc.com/2026/05/30/meta-struggled-selling-anything-other-than-ads-will-ai-be-different.html). Meta still counts on ads for 98% of revenue. Schackart said he wants to see "tangible evidence of a growing list of new, AI-first products created by Muse Spark, even if monetization lags." He said that's "what investors are looking for." No matter how good Wang's model may be, Zuckerberg has a high hill to climb with developers coming off the Llama debacle. "I think the AI community largely ignores Meta at this point," said Rob May, CEO of the startup Neurometric, which works in the realm of token engineering.... Krish Subramanian, the CEO of consulting firm KOI AI and former product head at IBM Consulting, said developers are more excited about Google's AI models than what Meta is offering. The appeal of Llama was that it specifically targeted developers wanting open-weight alternative models, while with Muse Spark, Meta has made little effort in that direction, he said. "The lack of developer trust will come back to hit them if they don't focus on third-party developers," Subramanian said, noting that it took years for Microsoft to regain trust from open-source coders during the early days of Azure. "To just focus on a walled-garden kind of an ecosystem and ad revenue as the main source of income, they probably will never become the big player," he said. A Meta spokesperson pointed to Wang's recent comments about the company's continued support for the open-source ecosystem, and said Meta still plans to offer outside developers access to Muse Spark's underlying technology via an API, as it previously announced. "We're already testing with some early partners, and look forward to releasing it this month," the spokesperson said. _ "That Zuckerberg's [metaverse and virtual reality ambitions](https://www.cnbc.com/2026/01/13/meta-lays-off-vr-employees-underscoring-zuckerbergs-pivot-to-ai.html) have generated [over $80 billion](https://www.cnbc.com/2026/04/29/metas-reality-labs-lost-over-4-billion-in-first-quarter.html) in total losses since late 2020 makes the AI pitch a tougher sell," the article points out, citing this observation from Howard Yu, business professor at Switzerland's International Institute for Management Development. "He's running out of the space for his credibility to last," Yu said. "I think the virtual reality foray may have burned up a lot of his goodwill in front of investors."

 Meta is planning to launch its own prediction market app to compete with companies like Kalshi and Polymarket in a booming sector that some analysts project could become a [$1 trillion industry](https://www.cnbc.com/2025/12/17/prediction-markets-trillion-dollar-trading-volume-ek-report.html) in the coming years. Meta CEO Mark Zuckerberg has instructed a team to start building a standalone app called Arena where people can guess the outcome of real-world events, according to two employees who were not authorized to speak publicly about the project. Meta's plans to build a prediction market app were [first reported](https://www.nytimes.com/2026/06/23/technology/meta-prediction-markets-app.html) by _The New York Times_. NPR obtained internal Meta documents about the effort that have not been previously reported. The documents describe how instead of wagering real money, users of the new app will receive "a daily virtual allotment" of "play money" that can be used to bet on "the outcome of future events." On the most popular prediction market apps, Kalshi and Polymarket, traders bet billions of dollars a week on everything from what [Rotten Tomatoes score](https://kalshi.com/markets/kxrt/rotten-tomatoes-scores/kxrt-supe?op_market_ticker=KXRT-SUPE-60) a forthcoming movie will receive to [when Israel](https://polymarket.com/event/usisrael-strikes-iran-by) will drop bombs on Iran. While it is unclear what range of topics Meta hopes to cover with its prediction market, the company is planning to use artificial intelligence to drive the service, according to the internal documents. The new app, which has been codenamed "Antwerp" and "FBForecast," will use Llama, the company's large language model, to automatically generate questions from trending topics. Meta's AI will also make "personalized market recommendations" to users who download the app, according to the documents. That same artificial intelligence will resolve markets, according to the documents. In other words, AI will have the final say over whether something did or did not happen. As on Kalshi and Polymarket, the wagering happens around "yes" or "no" questions. This AI-driven resolving process will happen in "near real-time," the documents state. In 2020, Meta released an app called Forecast, a crowdsourced prediction market app where people could guess about what might happen in the world, including predictions about the course of the pandemic. The app was wound down two years later. The internal documents NPR reviewed cited "the operational cost of manual question curation" as the reason Meta shut that effort down. The documents say the new app now under development is a "rebuild" of that app. When a prototype app is finished, Meta employees will test it before its full public launch for both iPhones and Android devices, the documents state. An exact release date was not specified. Meta declined to comment. Developing a prediction market without real monetary stakes may seem counterintuitive, since the hope of making a handsome profit has drawn millions of users to Kalshi and Polymarket. But gaming lawyer Daniel Wallach, who watches the prediction market industry closely, said releasing the app without the ability to wager money could allow Meta more time to apply for a license to operate from federal regulators and provide breathing room for the contested legal landscape to become more settled. "We're clearly in legal limbo," said Wallach, noting the more than 30 [pending lawsuits](https://www.npr.org/2026/01/30/nx-s1-5691837/lawsets-prediction-market-kalshi) over the legality of prediction markets. "And we might not have a clear answer for another year or two." Meta moving forward with its own prediction market service makes it the biggest-yet Silicon Valley player to enter the fray, even as questions over [market manipulation](https://www.npr.org/2026/06/02/nx-s1-5843371/george-santos-kalshi-insider-trading-investigation), [insider trading](https://www.npr.org/2026/05/27/nx-s1-5836659/google-polymarket-insider-trading-doj-charges) and how the markets can reshape reality or create perverse incentives continue to plague the sector. Nonetheless, the industry has the backing of the Trump administration, with officials beginning the process of rewriting the federal rules overseeing prediction markets, which were shunned for decades by regulators. "The prediction markets of today are a novel creation that only arrived at the dawn of the second Trump administration," Wallach said. "And it's launched all these legal fights over whether Congress ever intended for this kind of activity to be regulated by the CFTC," he said, referring to the Commodity Futures Trading Commission, the agency overseeing prediction market sites. Meta is just the latest company hoping to tap into the surging popularity of prediction markets. Traditional sports gambling sites, like [DraftKings](https://www.draftkings.com/draftkings-debuts-predictions-app-entering-prediction-markets) and [FanDuel](http://linkedin.com/posts/amy-howe-12386431_today-were-launching-fanduel-predicts-activity-7408921056821833730-TO3l/), have released their own prediction market services. President Trump's TruthSocial has [announced](https://blockworks.com/news/trump-media-truth-predict) prediction market plans. And more than 70 other companies have launched prediction market projects or services, [according to Event Horizon](https://nexteventhorizon.substack.com/p/introducing-the-prediction-market-startup-database), a newsletter that tracks the prediction market industry. More than 3 billion people use at least one Meta app every day, including Facebook, Instagram, WhatsApp and Threads. But the company has struggled to get new standalone apps off the ground.

184920000 story [](//tech.slashdot.org/index2.pl?fhfilter=facebook)[ ](//tech.slashdot.org/index2.pl?fhfilter=opensource)[](//tech.slashdot.org/index2.pl?fhfilter=meta) Posted by [BeauHD](https://www.linkedin.com/in/beauhd/) on Monday August 10, 2026 @02:00PM from the no-cloud-required dept. Meta has [released](https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model) Muse Glimmer, a slimmed-down open-weight AI model [designed to run locally on a single GPU](https://www.engadget.com/2233312/metas-open-source-muse-glimmer-model-can-run-on-a-single-computer/) for agent tasks such as scheduling, file management, coding, and tool use. The release is based on Meta's closed Muse Spark 1.2 model and appears to be aimed at attracting developers who want capable AI agents without relying entirely on cloud-hosted services. Engadget reports: _Facebook said that it's making the "weights" that AI systems use to choose responses available to everyone [on Hugging Face](https://huggingface.co/meta-models/Muse-Glimmer-30B) along with developer documentation. The download is available for free, and users can run the model on their own PCs. The company noted that optimized integrations will land on llama.cpp and other sites, "so you can go from download to working agent in minutes." The model is powerful for its size, according to Meta, with the "strong success rates" on benchmarks like DeepSearch QA, MCP-Atlas and SWE-Bench (which evaluates its ability write and debug code). It also supports reliable tool use, multi-step reasoning, failure recovery, multimodal input and scaffold compatibility for work with OpenClaw and other agent orchestrators. It was trained on data from over 100 languages, the company added. _"Rather than centralizing superintelligence, we should distribute it widely and give every person the ability to direct it," CEO Mark Zuckerberg said in [an essay](https://about.fb.com/news/2026/08/the-future-is-for-everyone/) accompanying Muse Glimmer's release. "This has the potential to begin a new era of personal empowerment where individuals can use this powerful new capability to reach their full potential, pursue their interests, and improve their lives and the world more than ever before."