Debate is growing over who should bear legal responsibility when AI systems cause harm, as recent incidents involving AI models from OpenAI and Anthropic revived the question. Some prominent figures argue AI companies should be liable under existing consumer protection laws. Legal actions and political calls are increasing, but experts warn that applying intent-based legal standards to autonomous AI may be difficult, especially for open-source models and inherently unpredictable behaviour.

Companies are deploying AI agents at work, but employees say the tools often demand ongoing monitoring, retraining and result reviews, increasing management burdens. A product chief at a British social impact investment platform said a customer support agent drifted beyond its remit and fueled complaints. A Boston Consulting Group report put AI-agent integration at 30 percent, up from 13 percent. BCG described “AI brain fry” from excessive oversight, while others cited compliance, accountability and performance degradation risks.