The Food and Drug Administration launched Elsa, an artificial intelligence assistant designed to accelerate drug and medical device approval processes. Agency officials intended the generative tool to eliminate bureaucratic bottlenecks while reducing administrative burdens for staff members. The platform promised revolutionary improvements for reviewing life-saving treatments through automated workflow management. However, implementation has revealed significant operational flaws that threaten public safety. The technology was positioned as a major advancement for regulatory efficiency.
Current and former FDA employees report that Elsa frequently produces false information and creates nonexistent clinical research references. The system generates misleading summaries of actual studies and fabricates entire research papers that cannot be verified. Staff members describe disturbing patterns where the AI distorts legitimate scientific findings. Agency leadership has demonstrated minimal concern about these documented problems. Workers face no mandatory training requirements for proper tool usage.
The core problem involves insufficient human supervision rather than technological limitations alone. Voluntary implementation policies create dangerous gaps in oversight for critical health decisions.
Current and former FDA employees report that Elsa frequently produces false information and creates nonexistent clinical research references. The system generates misleading summaries of actual studies and fabricates entire research papers that cannot be verified. Staff members describe disturbing patterns where the AI distorts legitimate scientific findings. Agency leadership has demonstrated minimal concern about these documented problems. Workers face no mandatory training requirements for proper tool usage.
The core problem involves insufficient human supervision rather than technological limitations alone. Voluntary implementation policies create dangerous gaps in oversight for critical health decisions.
