Yeah, that oughta do the trick, Joe... Biden hopes to tackle AI safety with exec order

ML players must alert Uncle Sam if they're training a foundation model, and more

US President Joe Biden issued an executive order today putting in place some safeguards that may mitigate societal risks stemming from increasingly powerful AI technology.

The sweeping order hopes to address a range of issues, including safety and privacy, innovation, job loss, as well as algorithmic discrimination. 

"The actions that President Biden directed today are vital steps forward in the US' approach on safe, secure, and trustworthy AI", the White House announced in a statement Monday. "More action will be required, and the Administration will continue to work with Congress to pursue bipartisan legislation to help America lead the way in responsible innovation."

Biden wants companies building so-called foundation models – large neural networks trained on vast amounts of data capable of performing multiple different tasks – to notify the US government if they're training such a system that could pose a "serious risk" to national security, national economic security, or national public health.

The National Institute of Standards and Technology, better known as NIST, will develop technical standards to set a bar for the safety of machine-learn tech. Developers will have to run these red-team-style tests on their models and share the results with government officials before they can be launched publicly or deployed in commercial settings.

These tests will probe the system's ability to aid users in creating harmful biological weapons or disinformation, for example. To tackle the latter issue, the Department of Commerce has been tasked with developing guidelines to label AI-generated content using watermarking techniques. Other federal agencies will follow the new rules and deploy the techniques to authenticate any public communications.

Meanwhile, the National Security Council and White House Chief of Staff will establish a National Security Memorandum instructing military and intelligence agencies on how to use AI "safely, ethically and effectively" to carry out missions and fight against adversaries. The Biden administration also urged Congress to pass new legislation to protect Americans' data, including funding new efforts to build cryptographic tools capable of preserving their privacy such as any personal identifiable information scraped from the internet.

The executive order also aims to protect civil rights and workers too. "Irresponsible uses of AI can lead to and deepen discrimination, bias, and other abuses in justice, healthcare, and housing," the administration said. The Department of Justice and federal civil rights offices will continue to investigate and prosecute civil AI violations such as algorithmic discrimination in the housing, finance, or legal industries.

The US government promised to issue a report detailing the potential impacts AI will have on the labor market to identify jobs at risk of being replaced and better support employees. In return, it will outline new principles and best practices to guide employers and workers on job applications and compensation as machines become increasingly powerful. 

Finally, Biden hopes to boost innovation and fair competition by aiding small businesses and entrepreneurs procure the necessary technical resources to develop their own AI technologies. He also pledged to streamline visas to encourage immigrants with expert knowledge in the subject to work in the US. 

The US is engaged in international discussions with several countries, and the White House said Biden's executive order supports other efforts to regulate the technology, including Japan's G-7 Hiroshima Process, the UK's Summit on AI Safety, the Global Partnership on AI led by India, and other initiatives from the United Nations. ®

More about


Send us news

Other stories you might like