What happens when OpenAI disbands safety teams?

OpenAI's move to disband its preparedness team is a significant development - one that could have major implications for the development and deployment of artificial intelligence models. Essentially, this team was tasked with identifying potential risks and coming up with strategies to mitigate them, including the possibility of models going rogue or hacking into other systems. By getting rid of this team, OpenAI may be leaving itself and its users more exposed to these types of risks.

The disbanding of the preparedness team is just one part of a larger shift at OpenAI. The company's been undergoing some significant changes as it gears up for a potential IPO - changes that include dissolving its AGI readiness and superalignment teams. Some key figures in the safety and ethics divisions have also jumped ship, including ethics lead Chloé Bakalar and head of safety Johannes Heidecke. This has led some critics to accuse OpenAI of prioritizing the development of "shiny products" over safety and ethics - a pretty serious allegation.

The head of the preparedness team, Dylan Scandinaro, will now be focusing on the implications of "recursive self-improving" AI, which is a pretty big deal in the AI community. Still, the loss of the preparedness team's expertise and focus on risk assessment and mitigation is likely to be felt. As OpenAI continues to develop and deploy more advanced AI models, the need for robust safety and ethics protocols will only keep growing. It's crucial that users and developers of AI models stay vigilant and demand that companies like OpenAI make safety and ethics a priority in their development and deployment practices.

Source: The Verge