AI adoption soars even as trust lags behind, says Dataiku’s AI Confessions report

Artificial intelligence is infiltrating corporate workflows at an unprecedented pace, but leaders are racing ahead even as trust struggles to keep up. According to the new Global AI Confessions report, conducted by Dataiku, AI agents are now embedded in daily operations at 86 per cent of large organisations, with nearly 42 per cent outsourcing dozens of internal processes to automated systems handling data movement, operational tasks, and even critical decision-making.

Despite this heavy reliance, confidence in AI remains patchy. Only 19 per cent of executives always demand explanations for AI-driven decisions, and trust drops dramatically when it comes to sensitive areas such as regulatory compliance, recruitment, and ethics, just 11 per cent believe AI is ready for these high-stakes arenas. Nearly 95 per cent of Chief Data Officers cannot provide full traceability to regulators, leaving organizations exposed in an increasingly scrutinized environment.

The report also shows that AI adoption is hitting real-world hurdles: 52 per cent of organisations have slowed their AI rollout, while 58 per cent of pilots fail to move beyond proof-of-concept. Operational bottlenecks caused by flawed AI decisions affected 59 per cent of teams last year, and 75 per cent of organisations cite lack of trust as the top barrier.

Adding fuel to the fire, 91 per cent of leaders acknowledge “shadow AI”, the use of unapproved AI tools, is thriving within their organizations, creating risks that governance frameworks struggle to manage.

Experts warn that the gap between AI adoption and organisational trust is widening, leaving companies caught in a delicate balance: innovate too slowly and fall behind; embrace AI recklessly and risk serious operational and ethical failures.