Diversity and inclusion quotas based on gender and race have been quietly rolled back in 2025. But as AI reshapes work, HR experts say a new kind of quota may emerge.
The demographic it aims to deliver equity for? The human race.
At its HR Symposium in London this Tuesday, analysts from research giant Gartner predicted that by 2032, at least 30% of the world’s top economies will require “certified human quotas mandating a minimum level of human involvement in work.”
Ania Krasniewska, group vice president of Gartner, told Business Insider during an interview at the conference that the global research firm’s prediction was based on several Gartner analyses.
The goal would be to ensure humans continue to play a meaningful role in production, decision-making, and creative processes even as AI plays a greater role in businesses’ workflow and output.
“This kind of change won’t be organizationally driven; it will be legislation driven,” said Krasniewska. “When these rules kick in, organizations will need clear processes to redeploy employees and a way to prove they’re doing it.”
In August, Australia’s High Court ruled that the country’s Fair Work Commission can investigate whether an employer could have redeployed workers by restructuring work before declaring a redundancy.
The decision arose from a case involving a business outsourcing to contractors while laying off employees, but it will also impact companies’ decisions about the redeployment of the human workforce alongside AI.
“We expect to see more legislation like this to be announced in the coming years,” Krasniewska said.
As more work is produced by AI, business leaders will also need to consider how to handle accountability for errors that AI-produced work may create.
Related stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Krasniewska pointed to the example of medical imaging: “If AI read it, you pursued a course of treatment, and it turned out that the reading was wrong, who takes accountability for that?”
A popular fix is to keep a “human in the loop.” The EU’s AI Act, for example, requires a “meaningful” level of human oversight for high-risk AI systems, ensuring that people can step in to correct AI outputs that endanger health, safety, or fundamental rights.
It’s not a foolproof solution.
This week, Deloitte, one of the world’s largest consulting and accounting firms, agreed to partially refund the Australian government after it produced a report with errors, including academic references to people who didn’t exist and a made-up quote from a Federal Court judgment. The Big Four firm said it used AI to help write the report.
Krasniewska said businesses need to consider how to track where the human interaction components are and use citations and watermarks on information so that it is clear whether it had an AI component or not.
There will likely be legislation or public pressure around the need to have some form of disclosure, she added.
“Organizations haven’t necessarily thought that far ahead as to what we should do. But I think there’s a very human reality of sensing that we will need to disclose how we got from point A to point B,” said Krasniewska.
