New research from Altimetrik, in partnership with HFS Research, finds that the vast majority of the world’s largest businesses have deployed AI without answering a fundamental question: Who is actually in charge of what it decides? 

And yet, AI is shaping hiring decisions, influencing capital allocation, triggering compliance actions and steering operational trade-offs at a scale no executive team can realistically supervise. 

The report, Humans at the Helm of AI, is based on a survey of more than 500 senior executives across Global 2000 organisations in five industries and reveals that only 14% have a documented AI strategy with clear goals. 

"AI is accelerating decisions across the enterprise, but, done well, it requires deep engineering discipline,” said Raj Sundaresan, CEO of Altimetrik. “Too many organisations are scaling AI without redesigning accountability, which risks scaling bad decisions faster. Putting humans at the helm is about ensuring every AI-driven decision is governed with the same engineering rigour, ownership, and scrutiny we expect from any critical business system. Without that accountability, you’re scaling risk instead of intelligence.” 

The report identifies a chasm between organisations that have institutionalised AI as a governed enterprise capability and those still running it as a collection of team-level experiments. Only 13% have reached high maturity, and they are more than twice as likely to report faster, more accurate decisions and measurable customer and revenue impact. Everyone else is stuck managing long execution cycles, ownership ambiguity and governance structures that were designed for a world without AI. 

“Enterprises are scaling AI faster than accountability, and that gap is now a workforce crisis,” said Phil Fersht, founder and Chief Analyst of HFS Research. “When leaders don't define what AI decides and what humans own, employees stop questioning it. That's not augmentation, it's abdication. Fix it now, or you're not building an intelligent organisation. You're scaling unmanaged risk." 

That pattern is visible across the survey. More than half (52%) of employees say fear of replacement is their biggest barrier to engaging with AI. Nearly 80% receive fewer than 10 hours of training per year. And the skill that AI oversight most depends on, the ability to challenge AI outputs, ranks last among the capabilities executives say they value. The danger with this approach is that it leaves businesses with a workforce that has learned to follow AI rather than govern it. In fact, 75% of organisations say their teams defer to external partners because they lack the confidence to push back. 

Read the full report for complete findings.