How is AI being used by state employees in Iowa? Senate bill seeks to define uses
Walter Bibikow via Getty Images
An Iowa Senate bill that advanced from a subcommittee Tuesday would limit the uses of artificial intelligence in state agencies.
This article was originally published by Iowa Capital Dispatch.
An Iowa Senate bill that advanced from a subcommittee Tuesday would limit the uses of artificial intelligence in state agencies.
Senate Study Bill 3014 would require state agencies to disclose the use of artificial intelligence tools and prohibit the use of artificial intelligence in matters that impacted a state employee’s employment status, bargaining rights, wages, transfer or eligible hours.
Sen. Charlie McClintock, R-Alburnett, said the bill is “setting some guidelines in the use of AI.”
“In short, this bill outlines things that basically humans do now, that AI has kind of reached into both now and into the future that could take over those practices,” McClintock said.
Technology groups, state agencies and business entities expressed interest in the bill, but only unions representing state employees registered in favor of the bill.
Peter Hird on behalf of the Iowa Federation of Labor, AFL-CIO, said in the subcommittee that the bill is “getting out in front” of artificial intelligence.
“These are human beings that work for the state of Iowa, we need to make sure that we’re treating them with the dignity and respect they have,” Hird said.
Reports show that artificial intelligence models are widely used by employers to screen resumes, script interviews or gather data about candidates.
Recently, an AI recruitment company, utilized by a slew of big name companies including Microsoft and PayPal, was sued by the state of California for allegedly compiling reports to screen candidates for hiring positions without their knowledge.
Hird said the technologies are being used by companies for other types of employee evaluation or automations that have eliminated jobs.
Tom Chapman, speaking on behalf of the Iowa Catholic Conference, said the group was leaning toward supporting the bill and noted that the challenge with AI is “not really so much technological, but anthropological.”
“It’s a matter of protecting human identity and authentic relationships,” Chapman said. “When we turn over decision making, perhaps, to AI, creativity, I think it risks diminishing our own agency, our own imagination and our own personal responsibility.”
Sen. Liz Bennett, D-Cedar Rapids, who served on the committee and signed off on the bill, said AI models can often misinterpret data, hallucinate or make faulty inferences, depending on how they are trained.
“End users don’t always understand this, so that really can produce a situation where a machine is limiting somebody’s opportunity, or making the decision to fire somebody, and I think we need to keep humans in the loop,” Bennett said.
Other groups commented that some of the broad definitions in the bill could limit a state agency’s ability to utilize some of the efficiencies provided by artificial intelligence.
Andrew Ritland, speaking on behalf of the Mahaska County attorney’s office, pointed to language in the bill that stipulated a state agency could not “transfer existing duties or functions performed by state agency employees to an artificial intelligence system.” Ritland said the language was “short sighted.”
“AI is a tool just like any other, and if there is a desire for government to be more efficient, run leaner and ultimately save the taxpayers money, using artificial intelligence systems to assume duties, maybe data compilation or other duties, is a labor saving measure,” Ritland said.
McClintock, who introduced the bill, said he would go ahead and move it forward, with the understanding that lawmakers would do some work to address the concerns raised before bringing it to the committee.
“AI is not going away,” McClintock said. “It’s only going to continue to be brought forth in technology… so we’ll just keep addressing it. We can improve this bill and we certainly will.”