Michelle Donelan, head of the Department for Science, Innovation and Technology, unveiled the funding programme during the second day of the AI Seoul Summit. The UK’s AI Safety Institute (AISI) will lead the new initiative, which it will deliver in collaboration with UK Research and Innovation (UKRI) and The Alan Turing Institute. Shahar Avin, an AI …
The U.K. has debuted what it calls a landmark toolset for artificial intelligence (AI) safety testing. The new offering, announced Friday (May 10) by the country’s AI Safety Institute, is dubbed “Inspect” and is a software library that lets testers such as startups, academics and AI developers to international governments assess specific capabilities of individual AI models and then generate …
Image Credits: Brain Every company or organization putting out an AI model has to make a decision on what, if any, boundaries to set on what it will and won’t discuss. Goody-2 takes this quest for ethics to an extreme by declining to talk about anything whatsoever. The chatbot is clearly a satire of what …
Once an AI model learns the tricks of deception it might be hard to retrain it. Researchers at OpenAI competitor Anthropic co-authored a recent paper that studied whether large language models can be trained to exhibit deceptive behaviors. They concluded that not only can a model learn to exhibit deceptive behavior, but once it does, …