Agentic AI Risks: Report Highlights Compliance Gaps

New Delhi, Aug 14 (PTI) As Agentic AI gains ground, about 86 per cent of enterprises surveyed anticipate heightened risks, though just 2 per cent of companies meet responsible AI gold standards, Infosys Knowledge Institute (IKI), the research arm of Indian IT giant, said in a report on Thursday.

The report, Responsible Enterprise AI in the Agentic Era, surveyed over 1,500 business executives and interviewed 40 senior decision-makers across Australia, France, Germany, the UK, the US, and New Zealand.

Put simply, agentic AI refers to Artificial Intelligence systems that have the ability to run autonomously, taking decisions and initiating actions with minimal human intervention.

The findings showed that while 78 per cent of companies see RAI as a business growth driver, only two per cent have adequate responsible AI (RAI) controls in place to safeguard against reputational risk and financial loss, according to a release outlining the major findings of the report.

Nearly 86 per cent of executives aware of agentic AI believe it will introduce new risks and compliance issues, it said.

The report analysed the effects of risks from poorly implemented AI, such as privacy violations, ethical violations, bias or discrimination, regulatory non-compliance, inaccurate or harmful predictions, among others.

It found that 77 per cent of organisations reported financial losses, and 53 per cent have suffered reputational impact from such AI-related incidents.

About 95 per cent of C-suite and director-level executives reported AI-related incidents in the past two years.

Nearly 39 per cent characterise the damage experienced from such AI issues as 'severe' or 'extremely severe', it said.

Highlighting the key findings, the report said AI risks are widespread and can be severe; responsible AI (RAI) capability is patchy and inefficient for most enterprises; and that executives view RAI as a growth driver.

With the scale of enterprise AI adoption far outpacing readiness, companies must urgently shift from treating RAI as a reactive compliance obligation to embracing it proactively as a strategic advantage, the report added.

To help organisations build scalable, trusted AI systems that fuel growth while mitigating risk, Infosys advocated studying practices of high-maturity RAI organisations who have already faced diverse incident types and developed robust governance; combining decentralised product innovation with centralised RAI guardrails and oversight; as well as embedding guardrails into secure AI platforms.