There's a near-even chance of a major AI lab reporting a model compromise by 2027.
The market indicates a slight lean towards the possibility of a major AI lab announcing a model compromise due to a cyberattack before 2027, with a probability of 54.30% for 'YES'. The confidence level of 60 suggests moderate uncertainty, and the market appears fairly priced with a small edge of -1.5.