The Compliance Digest by the Association of Governance Risk and Compliance (AGRC) has just published Oliver's latest article on "The AI Risk Equation: Delay vs Safety – Calculating the True Cost"
The article warns of the risks associated with hasty AI adoption. It highlights potential pitfalls like biased data, unpredictable behavior, and "hallucinations" (fabrications) in AI systems, illustrated through examples like the UK Post Office scandal, Air Canada's chatbot, and Zillow's failed real estate venture. It emphasizes the importance of human oversight, data quality, explainability, and ethical guidelines, as well as argues that delaying AI implementation is as dangerous as rushing into it, as competitors are already adopting it, and the companies who start now will get a significant edge in learning how to control this critical technology. The conclusion stresses the need for careful AI selection and robust safeguards to avoid catastrophic failures.