Infosys Knowledge Institute (IKI), the research arm of Infosys, has unveiled a critical report on the status of responsible AI (RAI) across enterprises in the "agentic AI" era. The report, "Responsible Enterprise AI in the Agentic Era," provides sweeping insights from a survey of over 1,500 business executives and 40 senior decision-makers across the UK, US, Australia, France, Germany, and New Zealand.
Despite 78% of companies recognising RAI as a growth driver, a mere 2% have put adequate controls in place to stave off reputational and financial risks. The potential damage from poorly implemented AI is evident, with 77% of organisations experiencing financial loss and 53% reporting reputational harm from AI incidents.
Key Findings:
- AI risks are pervasive: 95% of C-suite and director-level executives have reported AI incidents in the last two years, with 39% characterising these as "severe" or "extremely severe." A significant 86% of those aware of agentic AI foresee it introducing new risks.
- Inefficient RAI capabilities: Only 2% of companies meet the full standards of Infosys' RAI capability benchmark, "RAISE BAR." However, RAI leaders experience 39% lower financial losses and 18% less severity in AI incidents.
- RAI as a growth catalyst: Though seen as beneficial by senior leaders, companies believe they currently underinvest in RAI by 30%.
To bridge the gap in RAI readiness, Infosys suggests enterprises transition from a reactive compliance focus to proactive strategic adoption. Key recommendations include:
- Learning from RAI leaders: Analyse practices of high-maturity RAI organisations.
- Combining agility with governance: Balance innovation with centralised guardrails and oversight.
- Embedding security: Utilise platform-based environments to confine AI operations to pre-approved systems.
- Forming a proactive RAI office: Centralised monitoring and risk management enhance governance.
Balakrishna D.R., EVP at Infosys, emphasises the importance of establishing a responsible AI foundation resting on trust, data governance, and risk mitigation. Meanwhile, Jeff Kavanaugh asserts that robust RAI safeguards not only reduce risks but also unlock revenue opportunities as the "agentic AI" era takes hold.
As enterprise AI continues its accelerated adoption, prioritising RAI is imperative for sustainable growth and risk reduction.
