DENVER — Artificial intelligence is helping determine which Americans get job interviews, apartments, and even medical care, but the first major proposition dominated by bias in AI decision-making is that any facing headwinds from different directions.
Lawmakers working on these bills in states including Colorado, Connecticut and Texas will meet Thursday to make the case for their proposals, as civil rights groups and industry engage in a tug-of-war over the core parts of the bills.
Labor unions, consumer advocates and other groups are calling for greater corporate transparency and expanded legal avenues for public lawsuits over AI discrimination. The industry has expressed tentative support, but continues to pursue these accountability measures.
Bipartisan lawmakers caught in the middle, including those from Alaska, Georgia, and Virginia, have worked together to pass AI legislation in the face of federal inaction. The purpose of the press conference is to highlight efforts across states and stakeholders, and to emphasize the importance of cooperation and compromise in the first regulatory steps.
The senators include Democratic Sen. James Maloney of Connecticut, Democratic Senate Majority Leader Robert Rodriguez of Colorado, and Republican Sen. Shelley Hughes of Alaska.
“At this point, I don't have confidence that the federal government can pass anything quickly, and I do think there is a need for regulation,” Maloney said, adding that “industry advocates, government and academia It is important that we work together to create the best possible regulations and laws.”
Lawmakers say the bill is the first step toward future developments.
More than 400 AI-related bills are being considered in statehouses across the country this year, but most only target one industry or part of the technology, such as deepfakes used in elections or to create pornographic images. .
The largest bill introduced by this team of legislators would provide a broader framework for oversight, especially regarding one of technology's most perverse dilemmas: AI discrimination. Examples include AI that failed to accurately evaluate Black medical patients and AI that downgraded women's resumes when filtering job applications.
Still, up to 83% of employers use algorithms to assist with hiring, according to estimates from the Equal Employment Opportunity Commission.
If we don't do anything, there will almost always be bias in these AI systems, said Suresh, a computer and data science professor at Brown University who teaches a class on mitigating bias in the design of these algorithms. Venkatasubramanian explains.
“We have to do something tangible to stop people from being biased in the first place,” he says.
These proposals, primarily in Colorado and Connecticut, are complex, but their central argument is that companies would be required to perform “impact assessments” on certain AI systems. Those reports will include how AI will be incorporated into decision-making, the data collected, and analysis of the data. An analysis of the risks of discrimination and an explanation of the company's safety measures.
The main issue at issue is who can see those reports. Greater access to information about AI systems, such as impact assessments, means greater accountability and safety for the public. But companies are concerned that it also increases the risk of lawsuits and trade secrets being leaked.
Under bills in Colorado, Connecticut and California, companies would not have to submit periodic impact assessments to the government. Additionally, companies will be primarily responsible for disclosing to the attorney general if they discover discrimination. Governments and independent organizations do not test these AI systems for bias.
Unions and academics are concerned that relying too much on companies' self-reporting could jeopardize the ability of the public and governments to identify AI-based discrimination before it occurs.
“When you have a huge, multibillion-dollar corporation, it's already a big deal,” said Gersten Forseth, president of the Colorado AFL-CIO, a federation of labor unions opposed to the state's bill. Ta. “Essentially, you're giving them an extra bootleg to go after workers and consumers.”
Tech companies argue that increased transparency will reveal trade secrets in an increasingly competitive market. David Edmonson of TechNet, a bipartisan network of technology CEOs and senior executives pushing for AI legislation, said in a statement that his organization is working with lawmakers to “ensure that all legislation is implemented while allowing innovation to flourish. “We will ensure that AI risks are addressed.”
The California Chamber of Commerce opposes the state's bill, concerned that impact assessments could be made public in litigation.
Another controversial element of the bill is who can sue under the law, which the bill generally limits to state attorneys general and other public attorneys who are not citizens.
Workday, a financial and human resources software company, supported the proposal after a provision allowing residents to sue was removed from California's bill. Workday argues that civil lawsuits from citizens would be subject to judgment by judges, many of whom are not technology experts, and could lead to inconsistent approaches to regulation.
“It's clear that governments will have to intervene at some point because we can't stop AI from becoming part of our daily lives, but it's also obvious that the industry itself wants a good environment to thrive. ” said Vice President Chandler Morse. Workday Public Policy and Internal Affairs.
Sorel Friedler, a professor at Haverford University who focuses on AI bias, disagrees.
“It's common for American society to assert our rights by filing lawsuits,” Friedler said.
Sen. Maloney, R-Connecticut, said he and Rep. Giovanni Capriglione (R-Texas) were “authorized by industry,” despite industry spending millions of dollars lobbying against the bill. He said there was backlash against an article claiming that he was promoting the bill.
Maloney noted that one industry group, the Consumer Technology Association, is running ads and building websites urging lawmakers to reject the bill.
“We believe we are on the right path. We have been working with people from industry, academia and civil society,” he said.
“Everyone wants to feel safe and we are creating regulations that enable safe and reliable AI,” he added.
_____
Associated Press reporters Tran Nguyen contributed from Sacramento, California, Becky Boler from Juneau, Alaska, and Susan Haig from Hartford, Connecticut.
Bedayn is a corps member for the Associated Press/Report for America Statehouse News Initiative. Report for America is a nonprofit national service program that places journalists in local newsrooms to report on undercovered issues.
This article was generated from the Automata news agency feed without any modifications to the text.
