The Good, and the Gaps, in Biden’s AI Strategy

This article, edited for length and context, was originally published on LinkedIn.

Overview The Biden Administration is facing an inflection point as our government grapples with one of the most thrilling and challenging tech developments of our time, artificial intelligence (AI). On October 30, the president issued the United States’ first executive order focused on AI, an extensive 111-page EO aimed at ensuring "safe, secure, and trustworthy" AI development. At CTA, we’ve long said that the U.S. needs to lead the world on AI innovation – but is this EO the right way forward?

The Positive  

• The EO acknowledges AI's potential to address pressing challenges and enhance our prosperity, productivity, innovation and security. It aims to establish a coherent framework for AI, recognizing the need to avoid overlapping efforts among government agencies.  

• Additionally, the Executive Order emphasizes the importance of attracting global talent to fuel ongoing U.S. economic growth and competitiveness, proposing changes that would boost high-skilled immigration. 

 The Concerning  

• The EO’s support for expanded Federal Trade Commission regulatory authority over the AI ecosystem could invite unwarranted interference with AI innovators and the nascent AI market. Given the current FTC chair's antipathy toward innovation and American business, empowering the FTC as America’s default AI regulator would be a significant mistake. 

• The EO leans toward a top-down control approach over existing industry initiatives. Despite the fact that leading AI companies have voluntarily agreed to AI safety protocols, the administration invokes the Korean War-era Defense Production Act to compel companies to notify the government about training activities and results before releasing products to the public. This approach, in the words of technologist Adam Thierer, risks confining AI innovators in a "regulatory cage" that requires permission to innovate.  

• The EO opens the door to regulations affecting all facets of AI use in healthcare, research, drug and device safety, delivery and financing. When combined, these new regulations could hinder smaller companies from developing AI-led tools in healthcare. 

• The top-down approach in the EO also extends to the development of government standards for cutting-edge AI models. This contrasts with the voluntary industry standards model that has propelled American firms to leadership in nearly every global technology sector. Voluntary standards are crafted by industry stakeholders who deeply understand the market's needs and consumer demands, and allow rapid adaptation to changes in technology, consumer preferences and market conditions.