Welcome to the newest edition of ID Tech’s AI Update. Here’s the latest big news on the shifting landscape of AI and identity technology:
Meta is considering adding a facial recognition feature to its smart glasses. Dubbed “Name Tag”, the feature would automatically identify people spotted by a pair of glasses and pull up their information from Meta’s AI assistant, which would presumably be relying on info gathered through Meta platforms like Facebook and Instagram. Asked for comment, a Meta spokesperson tried to walk a tightrope in an effort to navigate privacy concerns, pointing out that it isn’t the first company to try this kind of thing: “While we frequently hear about the interest in this type of feature — and some products already exist in the market — we’re still thinking through options and will take a thoughtful approach if and before we roll anything out,” the spokesperson said.
ByteDance’s new Seedance 2.0 video generation model is being used to make lifelike, multi-shot video clips that are going viral on social media. It’s also causing some consternation due to an apparent lack of copyright guardrails, with clips featuring the likenesses of Hollywood stars like Will Smith, Brad Pitt, and Tom Cruise. Charles Rivkin, the head of the Motion Picture Association, issued a forceful statement condemning the tool’s launch. “In a single day, the Chinese AI service Seedance 2.0 has engaged in unauthorized use of U.S. copyrighted works on a massive scale,” he said. (Though it isn’t yet available outside of China.)
OpenAI has sent a memo warning the U.S. House ‘China committee’ that China-based DeepSeek is illicitly building its open source AI on American proprietary foundations. The memo told the U.S. House Select Committee on Strategic Competition between the U.S. and the Chinese Communist Party that DeepSeek continues to “free-ride” on frontier OpenAI models, essentially by programmatically analyzing ChatGPT outputs and using them to train its own models. OpenAI says this poses both business and safety risks if capabilities are copied without the corresponding safeguards.
Shanghai-based MiniMax has launched MiniMax-M2.5, pitching it as a frontier open weights model that can compete with the best, particularly when it comes to agentic tasks. It could be another DeepSeek sort of moment, given its relative efficiency. Processing 100 tokens per second, it can run continuously for one hour at a cost of just $1. “M2.5 is the first frontier model where users do not need to worry about cost, delivering on the promise of intelligence too cheap to meter,” the company said.
Anthropic has closed a $30 billion Series G funding round that values the company at $380 billion post-money. The round was led by GIC and Coatue, with participation from D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and MGX, among others. The OpenAI rival says its revenue run-rate has grown 10x every year for the last three years, and now stands at $14 billion, thanks in large part to its Claude Code system, with a bit of a boost from its Claude Cowork solution, which can actually take over a user’s browser and do things like make spreadsheets and slideshows.
A new class action lawsuit has been brought against xAI, alleging that its Grok AI chatbot sexually exploited women and girls by creating non-consensual sexualized deepfake images of them. Filed in a California federal court by a Jane Doe, the lawsuit aims to represent a subclass of South Carolina-based victims within the statute of limitations, and asserts that xAI is guilty of defamation, negligence, and the intentional infliction of emotional distress. It also says it violates California’s Unfair Competition Law. The plaintiff is seeking a jury trial.
The chatbot’s take: A little bit more explanation of what DeepSeek is allegedly up to, in OpenAI’s view:
–
By Alex Perala
