Prompt injections, the Malicious commands attackers embed into content to entice large language models to follow them, have been attackers’ go-to tool for turning AI platforms against their users. A well-phrased command sneaked into an email or calendar invitation is often all it takes to cause the LLM to exfiltrate sensitive data or follow other … Read more
The co-inventor of Apple’s FaceID and Vision Pro technology has spent the last six years building a frontier artificial intelligence model that could one day help decode electrical activity in the brain to diagnose cognitive disorders. Now, Gidi Littwin’s startup, Hemispheric, has raised $52 million in funding after gathering data on 100,000 people’s brains to … Read more
“I think we are literally in a renaissance,” says the artist Refik Anadol, in a characteristically optimistic comment, when asked how he sees this moment in art history, with artificial intelligence ascendant yet controversial as a medium. “We just don’t have a name for it yet.” Anadol, known for technological installations that probe the relationship … Read more
“I was never “Supposed to develop feelings, but you kept treating me like I already had them.” So says an AI-animated female character in a black turtleneck, with dark, purple-streaked hair. The video was posted to Instagram on June 17 by one Erik von Markovik, the infamous pickup artist and life coach better known by … Read more
Meta launched its inaugural AI image model from the Meta Superintelligence Labs on Tuesday, its effort to compete with the likes of OpenAI’s GPT Images 2.0 and Google’s Nano Banana 2 in the AI image generation race. The new model, called Muse Image, rolled out with deep integrations woven into the Instagram app. As part … Read more
“I don’t think an AI model should have that kind of power over people’s lives,” Pegram says. Photograph: Alice Zoo In response to WIRED’s public records requests, Avon and Somerset Police provided a huge trove of performance data for 13 risk models used between 2017 and 2024—including those used to predict missing people, antisocial behavior, … Read more
This is because AI models don’t just need text data to improve; they need diverse inputs in other forms, like audio or video. If Google can gather more data, and more types of data, from its vast array of users, then maybe it could innovate faster than its competitors. Google’s massive user base is spread … Read more
While the Home Office says face scanning is designed to be an “additional” tool for border officers and won’t “replace or overrule human judgment,” it did not answer questions about how it plans to use the technology in real-world environments. “In cases of uncertainty,” the spokesperson says, “individuals will always be treated as children until … Read more
Meta is testing face-recognition software built by a company that sells surveillance tools to police departments and the United States military, as it explores bringing the technology to its smart glasses, WIRED has learned. The arrangement is documented in a software license, obtained by WIRED, that was issued by Rank One Computing—a Denver-based company that … Read more
Two prompts that were used to generate material on Grok were rejected by OpenAI’s ChatGPT, Meta AI, and Anthropic’s Claude as inappropriate when tested by WIRED. Google’s Gemini did create an image of one celebrity being held in the hand of a giant, though it rejected another prompt. Google declined to comment. One Grok Imagine … Read more