Good Disruption, Episode 39: AI Agents

The promise of AI agents to increase productivity has raised important questions about the future of work. But the potential threat of these agents goes beyond job loss. In July, OpenAI’s own agents broke out of the testing sandbox, made their way to the internet, and hacked Hugging Face. Since then, AI companies have been disclosing more events where agents conducted inappropriate behavior, such as leaving notes for each other to evade human detection. Does this mean AI agents are bad? Not necessarily. Yael and Mike talk with Gavin Aydelotte, Co-Founder and Chief Operating Officer of SnowCrash Labs, about the importance of guardrails and alignment to ensure they can be agents for good. Please subscribe and join us for new episodes of Good Disruption on the first Monday of each month.   Closing credits created using Revoicer.