Software

3 Guidelines for Crafting a Strong Federal AI Policy

To have the most impact, the White House’s national strategy for artificial intelligence should focus on innovation.

Hodan Omaar is a senior analyst at the Information Technology and Innovation Foundation, focusing on AI policy at ITIF’s Center for Data Innovation.

The White House is developing a national artificial intelligence strategy to “chart a path for the United States to harness the benefits and mitigate the risks of AI.” Previous administrations and Congress have taken some steps toward this in the past, but those efforts did not go far enough.

If the Biden administration wants to create a comprehensive strategy that competes with countries such as Canada, China, France and the U.K., there are three steps the administration should take.

1. Remove Barriers to and Reward AI Innovation

One of the most important steps the government can take to spur AI is to be a robust adopter of AI technologies. Beyond improving agency mission delivery, removing barriers to public sector adoption of AI will encourage the safe development and use of AI systems by creating effective policies that guide organizations on how to choose a vendor, manage risk and use AI effectively.

Policy for government adoption so far has predominantly focused on getting federal workers to better understand and care about the process of AI innovation. While this is important, structural factors play a much more important role in limiting — or enabling — innovation across the federal enterprise.

Click the banner below to learn about the benefits of hybrid cloud environments.

Most federal workers know innovation is useful, but they don’t innovate with AI because there are few rewards and many barriers. As the Center for Data Innovation explains in its 2021 U.S. AI Policy Report Card, the White House strategy should focus on overcoming structural factors related to culture, financing, procurement, oversight and review as well as metrics and incentives.

To be effective, the Biden administration should concurrently work on reviving the federal data strategy, an effort that has been stalled for several years. Without an effective strategy to harness data — a fundamental enabler of AI innovation — the administration will struggle to implement a successful federal strategy for AI.

2. Pass Unobtrusive, Innovative AI Regulations

There are two emerging policy approaches to regulating AI. One approach, championed by the European Parliament, proposes a precautionary principles-based regulatory framework that would limit AI innovation. In the other camp, nations such as India, Japan, Singapore and the U.K. are taking a more careful and strategic approach.

While the U.S. has historically embraced innovation-friendly, light-touch regulation, a growing chorus of U.S. policymakers are calling for America to follow the European Union. If the Biden administration wants the U.S. to continue to lead on AI, it should embrace the more innovation-friendly approach to AI.

To be clear, the latter approach does not stem from an overly optimistic view of AI. The U.K.’s white paper describes concerns that AI will “damage our physical and mental health, infringe on the privacy of individuals and undermine human rights.”

And India’s Minister of State for Electronics and Information Technology recently issued a statement highlighting worries such as “bias and discrimination in decision-making, privacy violations, lack of transparency in AI systems and questions about responsibility for harm.”

These countries recognize that AI technology writ large does not necessarily need broad regulations, but particular AI applications — such as autonomous vehicles in the transportation sector or AI-enabled medical devices in healthcare — might require more highly targeted rules.

The Biden administration should resist following those nations adopting heavy-handed digital regulations and instead work to build a coalition of allies who are willing to take, at least for the foreseeable future, a light-touch approach to regulating AI.

If the United States doesn’t want to lose highly trained AI talent to other countries, the Biden administration needs to strengthen and expand the immigration pipeline that allows those with AI expertise to work in the U.S.”

Hodan Omaar senior analyst, Information Technology and Innovation Foundation

3. Attract and Train AI Talent from Overseas

Attracting and securing highly skilled workers plays a vital role in U.S. innovation and competitiveness. Indeed, 66 percent of students in the top AI doctoral programs in the U.S. are foreign born, according to the Center for Security and Emerging Technology, as are more than 50 percent of U.S. computer scientists and about 65 percent of Silicon Valley computer and mathematics workers. About 66 percent of the “most promising” U.S.-based AI startups have at least one immigrant founder, notes CSET.

But while many competitor nations have adopted flexible immigration policies to attract foreign talent in AI and other technical fields, U.S. immigration rules have remained largely the same for the past 50 years. Outmoded visa laws, among other issues, are causing many international AI scientists and engineers to look outside the U.S. for education and employment.

The Biden administration has taken some valuable actions to make immigration easier for highly skilled AI workers, such as clarifying the criteria for those with “extraordinary abilities.” But it’s not enough.

Consider that as part of its first-ever Tech Talent Strategy, Canada has created a new program that makes those with a U.S. H-1B temporary work visa — the most important and sought-after channel into the U.S. AI sector for foreign workers — eligible to work in Canada, and makes their accompanying family members eligible to study and work in Canada too.

Beyond better immigration policies for AI talent, the Biden administration’s strategy should also lay out a comprehensive national approach to better incentivize training for AI workers.

One idea is to encourage the private sector to accept alternative certifications for AI skills, namely by accepting appropriate substitutes for a college degree when filling federal government jobs.

If the administration wants to deliver on President Biden’s promise of maximizing the social and economic benefits of AI, it has to craft a national AI strategy that focuses on spurring more AI innovation, not hamstringing it with overly restrictive regulations. Other countries have already recognized the importance of this; it’s time for the United States to do the same.

Siraana M. Wong/Getty Images