Axios Event: Biden AI advisor says U.S. is “most pessimistic” country about AI

WASHINGTON, D.C. – Much of the current conversation and concern about AI-driven job displacement has centered on the argument that the technology will automate specific tasks within jobs rather than entire jobs themselves. However, experts at an Axios event also noted that the scale of automation will vary for different jobs.

• Axios tech policy reporter Maria Curi and business reporter Nathan Bomey moderated conversations with White House special advisor for AI Ben Buchanan, Rep. Ted Lieu (D-Calif.) and IBM Consulting global managing partner Jill Goldstein at the event, which was sponsored by KPMG.

Why it matters: Many workers are nervous and skeptical about how AI will impact or replace them, all while companies are figuring out their use cases for AI at vastly different paces.

What they're saying: "It is interesting that in global surveys the country that is most pessimistic about AI is the United States," Buchanan noted, echoing the concerns of many workers that exist alongside declining public trust in the technology.

"All of us use automation as part of our jobs right now. And it doesn't automate the whole job, it automates tasks within it. Which is not to say that there won't be jobs that have more or less automation, but I think we try to focus on the task level more than the job level," Buchanan said.

• Goldstein argued that there will be "extremes at both ends" for certain roles but said that "most people are in the middle."
• "There will be some roles where most of the tasks become obsolete, and there will be some roles that are newly emerged because of new skills that are required as a result of artificial intelligence," she said. "It's people who work with AI will replace people who don't work with AI," she continued.

There are also major concerns about AI's ability to perpetuate bias and discrimination in hiring. Buchanan called mitigating such bias a "huge priority" for the Biden administration, which recently released a set of principles for AI in the workplace.

• "What we've tried to do is put out, as directed by the executive order, [is] a series of best practices and principles from the Department of Labor, basically saying this is what it looks like to integrate well AI into the workplace … And if you don't do it well, here are some of the harms that we feel like could result," Buchanan said.

While tech industry culture has long been to "move fast and break things," Goldstein emphasized the ethical difficulties of adhering to that mantra in the development of AI. She cited concerns about recruiting bias in the early days of AI as a lesson "that sometimes, you have to go slow to go fast in a responsible way."

• "What happened was as an industry, we were going fast. And as a result, we were disadvantaging people in the process. So I think we learned from that pretty readily," she said.

When asked whether Congress needs to pass new laws to prevent AI-based discrimination in hiring processes or if it is a matter of enforcing existing laws, co-chair of the bipartisan AI task force Rep. Lieu outlined two potential approaches.

• One is to "simply fit AI" into an existing body of law, and another is that "we could also have laws that say you cannot deploy this system if it doesn't meet certain checks or balances" to reduce harm, Lieu said.

• "I think it remains to be seen how different states or the federal government approaches that," he continued.

Yes, but: Lieu said there is a low chance of Congress passing expansive legislation regulating AI use cases this year, saying that "in a Republican-controlled Congress, we're just trying to keep the lights on."

What we're watching: What makes it into the report Congress' bipartisan AI task force is tasked with creating to inform policy in 2025, and whether it will include guidance on AI in the workplace.

• Lieu said the task force is going through which bills and areas should make it into that report this month and that the report is still on track to be finalized by the end of the year.

Sponsored content:

In a View From the Top sponsored segment, KPMG Washington, D.C. office managing partner and U.S. federal government practice leader Patrick Ryan noted that creating secure environments for cloud and cyber are two of his biggest concerns for companies deploying AI.

• "Having access to a secure cloud environment to be able to deploy AI is really, really important … It's only going to become more important, particularly as I think about in this community here in D.C., I am really excited about the startup community," Ryan said.
• "The startup community is going to have a hard time accessing some of these tools and technologies if we don't make a widely available secure cloud environment available to them at a reasonable cost. It's available to the really big companies right now," he continued.