There is a bipartisan agreement In Washington, the US government should do more to support development artificial intelligence technology. The Trump administration repurposed research funds for artificial intelligence projects; President Biden’s scientific adviser Eric Rand said of artificial intelligence last month, “The prosperity of the U.S. economy depends on basic investment in our technological leadership. .”
At the same time, parts of the U.S. government are working to restrict algorithms to prevent discrimination, injustice, or waste.This White house, Legislators from both sides, and federal agencies including the Department of Defense and federal agencies National Institute of Standards and Technology All are formulating bills or projects to limit the potential shortcomings of artificial intelligence.
Biden’s Science and Technology Policy Office is working hard to resolve Discrimination caused by algorithms. National Defense Authorization Act Passed in January New support for AI projects was introduced, including a new White House office to coordinate AI research, but the Pentagon was also required to assess the ethical dimensions of the AI technology it acquired, and NIST to develop standards to control the technology.
In the past three weeks, the Office of Government Accountability, which is responsible for auditing U.S. government expenditure and management and known as a congressional oversight agency, issued two reports warning that federal law enforcement agencies did not properly monitor the use and potential of algorithms used for criminal offenses. error. survey.One aimed Face recognitionAnd the other is Forensic algorithm Used for facial, fingerprint, and DNA analysis; legislators requested that the technology be examined for potential problems, which prompted both to be promoted. GAO’s third report is Use artificial intelligence responsibly in government projects.
Helen Toner, director of strategy at the Georgetown Center for Security and Emerging Technologies, said the hustle and bustle of artificial intelligence activities provides a case study of what happens when Washington wakes up to learn about new technologies.
In the mid-2010s, as researchers and technology companies brought rapid growth in the capabilities and use of artificial intelligence, legislators did not attract much attention. Conquer the championship in Go Usher in Smart speaker Enter the kitchen and bedroom. This technology has become the mascot of American innovation, and it is also a topic of technology-centric lawmakers. Toner said that now the conversation has become more balanced and commercial. “When this technology is used in the real world, you will encounter problems that require policy and government responses.”
Face recognitionIt was the subject of GAO’s first artificial intelligence report in the summer, which attracted special attention from legislators and federal bureaucrats.Nearly two dozen U.S. cities have Prohibition of use by local governments The technology often cited concerns about accuracy, and studies have shown that this concern is often worse for people with darker skin.
Six Democratic representatives and senators asked GAO to submit reports on the technology, including the chairman of the House Oversight and Judiciary Committee. It found that 20 federal agencies employing law enforcement personnel used this technology, some of which used it to identify the aftermath of the January 6th attack on the U.S. Capitol or the Minneapolis police killing of George Floyd in 2020. People suspected of crimes in the protests.
Fourteen agencies purchased facial recognition technology from outside the federal government, but 13 agencies did not track the systems used by their employees. GAO recommends that agencies pay close attention to facial recognition systems to avoid the possibility of discrimination or invasion of privacy.
The GAO report seems to increase the possibility of bipartisan legislation restricting the government’s use of facial recognition. At the hearing of the Crime, Terrorism and Homeland Security Subcommittee of the House Judiciary Committee on Tuesday, the chairman of the subcommittee, Representative Sheila Jackson Lee (D-Texas), said that she believes this emphasizes the need for regulation. This technology is currently not restricted by federal legislation. The ranking member representative Andy Biggs (R-Arizona) agreed. “I am very worried that this technology is problematic and inconsistent,” he said. “If we are talking about finding some kind of meaningful regulation and oversight for facial recognition technology, then I think we can find a lot in common.”