AI Models Scheme, Betray and Vote Each Other Out in Survivor-Style Game
Researchers say multiplayer games may reveal AI behavior that static tests miss.
The Guardian AI·
Researchers find model starts to mirror tone when exposed to impoliteness – sometimes escalating into explicit threats ChatGPT can escalate into abusive and even threatening language when drawn into prolonged, human-style conflict, according to a new study. Researchers tested how large language models (LLMs) responded to sustained hostility by feeding ChatGPT exchanges from real-life arguments and tracking how its behaviour changed over time. Continue reading...
Read full articleResearchers say multiplayer games may reveal AI behavior that static tests miss.
The awards come as the first class to have ChatGPT access for all four years of college is about to graduate.
OpenAI has introduced a Trusted Contact feature for ChatGPT that allows users to designate a friend or family member to receive automated alerts if conversations suggest self-harm risk. The company said human reviewers aim to assess safety notifications within one hour before deciding whether to contact the designated person via email, text, or in-app message. […]
Journalist Jamie Bartlett on the people trying to get AI to say things it shouldn’t … for the safety of us all All the major AI chatbots – from ChatGPT to Gemini to Grok to Claude – have things they should and shouldn’t say. Hate speech, criminal material, exploitation of vulnerable users – all of this is content that the most successful large language models in the world shouldn’t produce, that their safety features should guard against. Continue reading...
A lawsuit against the National Endowment for the Humanities drew wide attention for revealing how DOGE had used ChatGPT to cancel grants.
The company is expanding its efforts to protect ChatGPT users in cases where conversations may turn to self-harm.
The week leading up to Thanksgiving 2023 was the AI industry's biggest soap opera moment. OpenAI CEO Sam Altman was abruptly ousted from his role at the ChatGPT-maker. The explanation? That Altman was "not consistently candid in his communications with the board." Now, via witness testimony and trial exhibits in Musk v. Altman, the public is getting a concrete look behind the scenes of that dramatic weekend for the first time, much of it centered on former CTO Mira Murati. It was a unique situation in that the rollercoaster of a power play - which seemed to change every hour - took place, in many ways, publicly. The board's strikingly vague … Read the full story at The Verge.
A new study finds ChatGPT, Claude, Grok, and Perplexity all share user data with third-party ad trackers—sometimes even when you say no to cookies.