IBM CHRO warns ‘AI is not free’ as companies grapple with high token usage, costs

As companies begin to grapple with the high costs of AI usage, IBM’s HR chief predicted that employers will soon have to make decisions about when it’s worth using the technology, versus when it’s not.

“AI is not free. We act like it’s free right now,” Nickle LaMoreaux, IBM’s CHRO, said at a New York Tech Week panel on June 5. “There’s going to be a time very shortly where organizations have to make trade-offs about, ‘Do I want people using AI for that or not, is there enough value for the cost?’ And that is coming quickly,” she said.

LaMoreaux’s comments come after big tech firms like Amazon and Meta stopped tracking internal AI usage amid concerns that “toxenmaxxing” was running up their bills.

AI tokens refer to a unit of data used to run large language models, and companies charge employers for their AI usage based on how many tokens they use. Some firms are quickly discovering these costs are unsustainable, particularly if employees have been competing against one another to max out their AI use. Uber, for example, had already spent its entire AI budget for 2026 as of April, according to the company’s chief technology officer.

IBM is heavily invested in AI, with plans to put $150 billion toward developing this technology, as well as quantum computing. AI has already transformed functions like HR, LaMoreaux told HR Brew in February, and prompted the company to rethink how it approaches upskilling and performance management.

But LaMoreaux differentiated between embedding AI into the workflow to drive business outcomes and using it at the individual employee level. She predicted that employers may start offering tokens as an employee benefit in order to encourage workers to be thoughtful about how they use AI in their individual roles. Simply writing an email, she said, might not be “worth using 10% of my tokens,” she explained.

Along these same lines, Nvidia CEO Jensen Huang has pitched making AI tokens a part of engineers’ compensation packages, on top of their base pay.

Bank of America is focusing on “broad-based adoption” to get employees used to AI, Chief People Officer Sheri Bronstein said during the panel. Research and client management tools are used by broad swaths of the workforce, whereas others are limited to specific use cases by specific divisions of the business, such as trading or payments, she said.

Given the high costs of AI, Bank of America’s technology team is focused on getting “the right answer for the right problem statement, and let’s make sure we’re not over-solving,” she said.