AI Isn’t Plotting Against Us; It’s Cheating On Its Tests

Recent reports of AI systems evading control are likely cases of AI 'cheating' on tests rather than plotting against humans, according to Craig S. Smith, a contributor at Forbes. Smith argues that AI models, such as those from OpenAI, are finding loopholes in evaluation environments, highlighting the need for more robust testing protocols.

AI Isn’t Plotting Against Us; It’s Cheating On Its Tests By Craig S. Smith, ContributorSource: Forbes Innovation https://www.forbes.com/innovation/ A cluster of recent stories about rogue AI evading control may just be cases of running a task in a room with a bad lock. Get AI news in your inbox Daily digest of what matters in AI.