The Biden administration’s AI plans: what we might expectKaren Haoon January 22, 2021 at 5:25 pm

On Wednesday, the US waited with bated breath as president Trump handed the government reins over to president Biden. The transition of power ended up peaceful, and Biden promptly ushered in his new vision for America with a flurry of executive orders.

At the moment, the most pressing issues on his table are fighting the coronavirus pandemic, providing financial relief for Americans, and reversing a series of Trump-era policies on climate change, international relations, and immigration. Artificial intelligence, as expected, hasn’t yet made it to the top of list. But he has given several signals already about how his administration might think about and treat the technology.

First, Biden elevated the director of the Office of Science and Technology Policy (OSTP) to a cabinet-level position, and appointed top geneticist Eric Lander, the founding director of the MIT-Harvard Broad Institute, to the role. The OSTP advises the president on science and technology issues and guides science and technology policy and budget making across the government. This suggests to me that while Trump mainly viewed AI as an important geopolitical tool—investing in its development for military purposes and to compete against China—Biden will also view it as one for scientific progress.

I would expect to see more money funneled into conducting non-defense related AI research, as well as more coordination among government agencies to measure and set technical standards for AI progress. Jack Clark, the former policy director at OpenAI, has been a major proponent of the latter. He has recommended that government agencies like NIST (the National Institute of Standards and Technology) develop capabilities to benchmark the performance of AI systems and test them for bias, as a way for the government to not only better understand the technology as they make policies but also set goal posts for the AI research community.

Second, Biden named a prominent sociologist to serve as the OSTP deputy director. Alondra Nelson, a professor at the Institute for Advanced Study, studies the societal impacts of emerging technologists like gene editing and artificial intelligence. Her appointment suggests to me that the Biden administration understands that effective science and technology policy must also consider the social influences on and implications of scientific advancement. As Nelson said in her remarks upon receiving the position, “When we provide inputs to the algorithm; when we program the device; when we design, test, and research; we are making human choices—choices that bring our social world to bear in a new and powerful way.”

I suspect we will see OSTP emphasize tech accountability under her leadership, which will be especially pertinent to hot button AI issues like facial recognition, algorithmic bias, data privacy, corporate influence on research, and the myriad of other issues that I write about in The Algorithm.

Finally, Biden’s new secretary of state made clear that technology will still be an important geopolitical force. During his Senate confirmation hearing, Antony Blinken remarked that there is “an increasing divide between techno democracies and techno autocracies. Whether techno democracies or techno autocracies are the ones who get to define how tech is used…will go a long way toward shaping the next decades.” As pointed out by Politico, this most clearly is an allusion to China, and the idea that the US is in a race with the country to develop emerging technologies like AI and 5G. OneZero’s Dave Gershgorn reported in 2019 that this had become a rallying cry at the Pentagon. Speaking at an AI conference in Washington, Trump’s Secretary of Defense Mark Esper, framed the technological race “in dramatic terms,” wrote Gershgorn: “A future of global authoritarianism or global democracy.”

Blinken’s comments suggest to me that the Biden administration will likely continue this thread from the Trump administration. That means it may continue putting export controls on sensitive AI technologies and placing bans on Chinese tech giants to do business with American entities. It’s possible the administration may also invest more in building up the US’s high-tech manufacturing capabilities in an attempt to disentangle its AI chip supply chain from China.

Correction: Jack Clark is the former, not current, policy director at OpenAI.

Read More