Last week, Biden signed an executive order establishing new standards for AI safety and security, demanding more transparency from tech companies developing AI tools. The order requires companies developing foundation models that pose national security risks to notify the government and share critical testing data. The move is seen as part of the US's effort to lead the way in AI regulation.
Key takeaways:
- Biden's cabinet has been exploring the capabilities of ChatGPT, a chatbot trained on a dataset of billions of parameters from various sources.
- A cabinet member asked ChatGPT if it could make a bioweapon, to which it responded that it couldn't assist with that.
- The Biden administration is investigating how AI models like ChatGPT are reshaping access to knowledge and how to regulate them without stifling innovation.
- Last week, Biden signed an executive order establishing new standards for AI safety and security, demanding more transparency from tech companies developing AI tools and requiring those creating models that pose national security risks to notify the government and share testing data.