As AI becomes more powerful, it faces a major trust problem. Consider 12 main concerns: misinformation, safety and security, the black box problem, ethical concerns, bias, instability, hallucinations in LLMs, unknown unknowns, potential job losses and social inequalities, environmental impact, industrial concentration and state overreach. Each of these questions is complex and difficult to resolve. But there is a consistent approach to closing the trust gap: train, empower, and include humans in managing AI tools.
The AI trust problem
