Suddenly, everyone wants to talk about how to regulate artificial intelligence

Recently, people’s attitudes towards artificial intelligence policy seem to have changed. For years, U.S. lawmakers and technology companies have been reluctant or even opposed to introducing strict technology regulations. Now both are asking for regulation.

Last week, OpenAI CEO Sam Altman appeared before a U.S. Senate committee to discuss the risks and potential of artificial intelligence language models. Altman and many senators have called for international standards for artificial intelligence. He called on the United States to establish an agency similar to the U.S. Food and Drug Administration to oversee artificial intelligence technology and urge its regulation.

For someone like me who is passionate about artificial intelligence policy, the Senate hearing was both encouraging and disappointing. It's heartening because the conversation seems to be moving away from empty self-restraints and toward setting regulations that might hold companies accountable. Frustratingly, this debate seems to have forgotten the last five-plus years of AI policy.

Picture | Sam Altman (Source: AP PHOTO/PATRICK SEMANSKY)

I'm not the only one who feels this way. “Suggesting that Congress start from scratch just feeds into the industry’s favorite narrative that Congress is so far behind and doesn’t understand technology — how can they regulate us?” Anna, a policy fellow at George Washington University’s Institute for Data Democracy and Policy Anna Lenhart said.

In fact, in the last Congress from January 2021 to January 2023, politicians proposed a large number of legislation on artificial intelligence. Lenhardt compiled all the AI ​​regulations proposed during this time.

They cover everything from risk assessment to transparency to data protection. None of them have reached the president's desk, but given that this trendy (or to many, "scary") new generative AI tool has captured the attention of Washington, Lenhardt hopes that some of them Some will be improved and reappear in new forms. Here are some issues to pay attention to.

Algorithmic Accountability Act

The bill was introduced by Democrats in the U.S. Congress and Senate in 2022, before ChatGPT existed, to address the real harms of automated decision-making systems, such as refusing to prescribe painkillers or denying their loan applications.

Lenhardt said the bill requires companies to conduct algorithmic impact and risk assessments. It would also put the Federal Trade Commission in charge of regulating and enforcing rules on artificial intelligence and increase its staff.

US Data Privacy Protection Act

This bipartisan bill seeks to regulate how companies collect and process data. Although an approach aimed at protecting the security of women's personal health data received widespread attention in the wake of Roe v. Wade, it was not passed in time. The debate surrounding the risks of generative AI may make it even more urgent than last time. The American Data Privacy and Protection Act (ADPPA) will prohibit generative artificial intelligence companies from collecting, processing or transferring data in a discriminatory manner. It also aims to give users more control over how companies use their data.

Artificial Intelligence Organization

Altman and several senators suggested setting up a new agency to regulate artificial intelligence during the hearing. But I think that might be a red herring. Lenhardt said the U.S. government needs more technical expertise and resources to regulate the technology, whether in a new agency or a revamped existing one. Having the power to enforce the law is a key requirement for any regulator, young or experienced.

Lenhardt said: “It’s easy to create an agency and not give it any powers. Democrats have tried to push the Digital Platforms Commission Act, the Data Protection Act and the Online Privacy Act to create new protections. Most have not. The USA Act, which enjoys bipartisan support, is doomed to fail, so none of these attempts will come to fruition.

Next?

Another tech-focused agency may be on the horizon. Senators Lindsey Graham, a Republican, and Elizabeth Warren, a Democrat, are working together to create a new digital regulator that would have the power to regulate social media companies and perhaps impose Social media empowers.

Democrat Chuck Schumer is also convening the Senate to introduce a new bill to specifically address the dangers of artificial intelligence. He received bipartisan support to pass a comprehensive artificial intelligence bill aimed at establishing safeguards and promoting responsible artificial intelligence development. For example, companies may be required to allow outside experts to audit their technology before its release and provide more information about its AI systems to users and governments.

Lenhardt said that while Altman appears to have won support from the Senate Judiciary Committee, the leaders of the House and Senate Commerce Committees will need to come together to support a comprehensive approach to AI regulation before it becomes law.

Regulations need to be put in place quickly before people lose interest in generative AI. “It’s going to be tricky, but anything is possible,” Lenhardt said.

Support: Ren

Original text:

https://www.technologyreview.com/2023/05/23/1073526/suddenly-everyone-wants-to-talk-about-how-to-regulate-ai/

The above is the detailed content of Suddenly, everyone wants to talk about how to regulate artificial intelligence. For more information, please follow other related articles on the PHP Chinese website!