The Morning Download: America’s HR Leaders Say We’re Thinking About AI Agents All Wrong

Good morning. There’s a lot of talk about AI as digital employees who contribute to hybrid teams alongside human labor, or “people” as we like to call ourselves. Not everyone is happy with that mindset.

Human resources leaders at some of America’s largest companies say it’s time to stop treating their AI as people, complete with names, job titles and key performance indicators.

Executives from IBM, Microsoft and other companies say thinking of AI agents as analogous to human workers hinders the effort to get full value from technology, the WSJ Leadership Institute’s Isabelle Bousquette writes.

Goodbye to all that. IBM used to have a series of agents that went by names like Harry, Hermione, Charlie and Sherlock. But it fell into a trap of focusing too much on each agent’s individual use cases rather than using them for more impactful large-scale process re-engineering.

“We learned this the hard way,” said Nickle LaMoreaux, chief human resources officer at IBM said, speaking Thursday at the WSJ Leadership Institute’s Chief People Officer Summit in Menlo Park, Calif.

“Too many CPOs are getting so hung up on: What does this agent do, what does this AI do?” she said. The biggest bang for your buck, she said, isn’t in individual assistant-type agents that, say, help write emails. It’s in integrating AI into enterprise workflows.

Some companies continue to push forward with the idea that AI agents should be considered fully formed “digital workers.” BNY, for example, said it employs dozens of AI “digital employees” that have company logins and human managers. They are built on the company’s proprietary AI platform Eliza, which is named after the wife of BNY founder Alexander Hamilton.

Automating tasks, not jobs. Microsoft’s agents do have names, albeit without a Broadway musical tie-in, said Chief People Officer Amy Coleman, also speaking at the summit. But it ends there.

“I don’t really think about roles and jobs that can be automated as much as I think about tasks in a job that can be automated,” she said.

Coleman also said she doesn’t envision a future where an AI agent is ever managing a human. Although, she added that things could change.

The shift away from equating AI agents with human workers has less to do with quelling human anxiety over layoffs and job replacement. It’s more a candid assessment of what the technology can actually do. Agents simply aren’t humans.

Even extremely capable, long-running agents can’t be held liable for their actions, said Box co-founder and CEO Aaron Levie, also speaking at the summit. “Accountability has to lie with humans. All our laws are set up to require that.”

That said, Levie added he believes that the transformation agents will drive inside organizations will be several orders of magnitude larger than previous waves of automation.

“This is the biggest shift we’ve ever seen in corporate work,” Levie said.

Do you manage AI agents like you manage people? And do they feel left out when the humans go out after work? Let us know.

Meanwhile...there's no joy in the corporate office. As companies race to find inefficiencies and cut costs, they're touting AI as the answer—and employees are feeling the squeeze.The WSJ's Mark Maurer and Chip Cutter take the temperature of corporate America. Perks from offsites to travel are vanishing, and many workers say AI feels less like a tool to help them work than a tool for extracting more output from fewer people. "There's almost nobody who is feeling positive vibes about their job right now," said Rocco Seyboth, a longtime software marketer outside Seattle. "We're in the AI dread era."

Recruiters make interviews ‘AI-free zones.’ While AI-generated applications overwhelm recruiters, companies like L'Oréal and EY are going old school, prioritizing in-person, AI-free interviews, FT reports. Others are throwing up new challenges, including video submissions and skills showcases to filter out low-commitment applicants. To be fair, many job seekers say they  turned to AI partly in response to its use by recruiters. 

Private credit’s exposure to the ailing software industry is bigger than advertised. Major private-credit funds run by Apollo Global Management, Ares Management, Blackstone and Blue Owl Capital are carrying significantly more software exposure than their filings indicate, a Wall Street Journal analysis found. On average, the four funds reported about 19% software exposure but actually held around 25%.The discrepancies stem from inconsistent categorization methods: Funds often slot software companies into healthcare, business services or other buckets. 

The AI series where sexy fruits cheat. A new viral dating show featuring sexy, cheating fruit could be proof that AI content can actually captivate viewers, the WSJ Leadership Institute's Isabelle Bousquette reports. Loosely based on the original “Love Island,” reality series, “Fruit Love Island” debuted earlier this month on TikTok, and has averaged more than 10 million views on each of its first 21 episodes.

“Visually, it’s very AI slop,” said Jaskaran Singh, who writes a weekly marketing newsletter. But, he added, “Story-structure wise, I don’t think it’s slop.”

Not everyone at OpenAI bought into Sora. The Journal looks at the rise and fall of OpenAI's video-generation tool, spotlighting some of the divisions within the AI startup. From the start, Sora's team operated as a closely guarded "startup within a startup," walled off from the core research group. Colleagues questioned the project's heavy compute allocation given its minimal revenue contribution. Others raised safety concerns about user-generated AI video.

Financial institutions continue to see promise in the European AI startup. French AI startup Mistral AI said it had raised $830 million from a consortium of banks in its first debt-financing operation, securing financial backing to run a new data center that will be powered by Nvidia chips, the WSJ reports. 

This tech exec talks politics. At a moment when much of the tech industry would rather not talk politics, Google executive Jeff Dean, chief scientist at Google’s DeepMind AI lab and a pioneer of the company’s artificial-intelligence strategy, has been a rare outspoken voice, the Journal reports.

President Trump is weighing a military operation to extract nearly 1,000 pounds of uranium from Iran, according to U.S. officials, a complex and risky mission that would likely put American forces inside the country for days or longer. (WSJ)

Transportation Security Administration workers may soon be paid. But airport security checkpoints are still short-staffed and lines are still snaking through terminals. Nearly 2,900 TSA officers called out from work on Saturday, or more than 10%, the Department of Homeland Security said Sunday. (WSJ)

Businesses and consumers across the eurozone grew more pessimistic about their prospects in March as the war in Iran is set to weaken economic growth, while firms reported that they expect to raise their prices at a faster pace to cover increasing costs. (WSJ)