AI Pioneer Explains Why Half Her Agents Failed
Sol Rashidi, who helped launch IBM's Watson, reveals why half her AI agents were terminated and what companies still get wrong about scaling AI.
Sol Rashidi, instrumental in launching IBM's Watson and a former chief AI officer, has dismissed half of her AI agent staff due to unreliability and unpredictability.
Rashidi stated that managing and correcting the AI agents took more time than training early-career human employees. She noted that while building agents might be feasible for small teams, at an enterprise level, only about one in three agents performs consistently well.
This gap between the hype and the reality of AI implementation is a significant concern for Rashidi. She warns that scaling AI predictably and safely remains a challenging endeavor, despite marketing suggestions of ease.
Referencing the early days of AI, Rashidi cautions against a premature reliance on these tools, comparing the current uncontrolled adoption to 'running with scissors.' She emphasizes that the original intent of AI was not solely replacement but augmentation.