OpenAI wins gold at prestigious math competition
Digest more
Chain-of-thought monitorability could improve generative AI safety by assessing how models come to their conclusions and spotting the "intent to misbehave.”
OpenAI recently revealed to Axios that its AI chatbot now handles more than 2.5 billion user prompts per day worldwide, a remarkable leap from just 1 billion daily queries in December 2024. And of that figure, around 330 million prompts come from U.S.-based users alone.
The most powerful artificial intelligence company in the world just admitted it needs help from one of its biggest rivals to stay afloat.
The front page of the Google Store currently features a short video teasing one of the Pixel 10 smartphones it is announcing on August 20.
OpenAI has included Alphabet's Google Cloud among its suppliers to meet escalating demands for computing capacity, according to an updated list published on the ChatGPT maker's website.
OpenAI, the firm behind ChatGPT, has signed a deal to use artificial intelligence to increase productivity in the UK's public services, the government has announced.
OpenAI will reportedly release its own web browser, putting it in position to challenge Google’s long dominance in web search.
The researchers argue that CoT monitoring can help researchers detect when models begin to exploit flaws in their training, manipulate data, or fall victim to malicious user manipulation. Any issues that are found can then either be “blocked, or replaced with safer actions, or reviewed in more depth.”