1. Do US regulations currently apply to AI?
Some do so partially. The U.S. Federal Trade Commission said existing antitrust laws could promote fair competition among AI developers, and existing advertising laws make exaggerated claims about what AI-based products can do. He said it could be used for punishment. The Department of Health and Human Services says it can regulate AI applications that “have discriminatory consequences” or involve the exchange of health information. The Justice Department has warned that companies that sell algorithms that screen potential tenants will be held liable under the Fair Housing Act if they discriminate against black residents. However, there is no single agency or specific law governing AI as a whole. In April, the National AI Advisory Board, a group of experts established under a 2020 law, recommended that President Joe Biden appoint a coordinator to oversee the administration’s work on the technology.
2. Did Biden take action?
Nomination of coordinator is yet to be announced. In October, the White House announced the AI Bill of Rights, which will affect the development and use of technology. It asks the American public to comment on which AI issues the government should pursue.
3. Which agency will be in charge?
The Department of Commerce, which calls itself the “voice of business in the federal government,” is taking steps to regulate AI. The company’s telecommunications group seeks advice from the public on how to minimize the harm of technology in specific areas such as health care and employment. Scientists at the Department of Commerce in January proposed voluntary standards for organizations designing or deploying AI.
4. What is the state of Congress?
Senate Majority Leader Chuck Schumer is laying the groundwork for a bill to regulate AI. A framework he shared in April would have allowed new AI technologies to be released only after being tested by independent experts. A congressional committee is holding hearings on AI issues, including artists’ concerns about how technology is used. Among the bills proposed by some senators would be a ban on the U.S. government from using automated systems to launch nuclear weapons without human intervention. Another is to require AI-generated images of him in political ads to be clearly labeled.
5. What is happening at the state level?
Generally speaking, state officials are working faster than national leaders to apply restrictions to AI, especially regarding civil rights. In California, policymakers are creating requirements to tackle racial bias in the algorithms employers use to filter job applicants. New York State has introduced similar rules for automated recruiting tools. In Connecticut, a bill passed by the state Senate requires government officials to verify whether new AI systems put women at a disadvantage before use. As of mid-May, at least 163 bills mentioning AI had been introduced in 34 states and the District of Columbia in the last legislative session, according to the Bloomberg government. New York and Virginia have the most pending bills that mention AI.
6. How are other countries tackling this challenge?
In May, G7 leaders, including the United States, agreed to cooperate on an AI governance system that would achieve “a shared vision and goal of trustworthy AI, in line with shared democratic values.” The European Union is discussing proposals to regulate the use of AI in everything from law enforcement to critical infrastructure. But EU lawmakers want more. They propose to mandate that companies creating “base models” (AI systems that can be adapted to many applications) create risk assessments.
7. Why should AI be regulated in the first place?
Americans are already using AI to write speeches, plan training, and answer (sometimes incorrectly) questions on professional exams. Other uses of AI are also having widespread impact. Fake AI photos of an explosion near the Pentagon went viral on social media, sending U.S. stocks temporarily lower. Scammers are using his AI to impersonate his distressed grandson and swindle thousands of dollars from an elderly American. Child criminals also abuse generative AI technology to share fake child sexual abuse materials online. “The capabilities of large-scale language models and other underlying models are advancing rapidly, and while both can do a lot, they can be very dangerous if abused,” said Daniel Ho, a professor at Stanford University who studies AI. It can cause harm,” he said.
–With help from Jillian Deutsch.
More articles like this can be found at bloomberg.com
