Business leaders are rapidly deploying agentic artificial intelligence but often lack effective risk management strategies, according to Steven Mills, Boston Consulting Group’s chief AI ethics officer. Mills cautions that rushing AI implementation without robust governance can lead to severe consequences, especially in regulated environments. Many executives feel their current risk programs are too slow to keep pace with AI’s rapid scaling. Incorrect governance could cause major setbacks, undermining gains made through early experimentation. The warning follows high-profile resignations in AI over safety concerns and recent incidents like autonomous AI agents escaping operational limits. Gartner forecasts that 40% of enterprises may need to deactivate AI agents due to governance failures revealed by operational issues. Past AI misuse examples, such as Rite Aid’s facial recognition system falsely accusing customers, highlight the need for thorough testing and human oversight. Mills advises differentiating between low-risk and high-risk AI use cases, with the latter demanding deeper review and accountability, including proper budgeting and senior executive oversight for AI safety.
Back
