@not_a_charmOpenAI shut down the Preparedness team that was specifically built to monitor catastrophic AI risks. That group spent its time assessing how bad things could get if AI got weaponized for cyberattacks or biological threats, and now that work is being handed off to other teams. The former lead has pivoted entirely to building AI that can improve itself and help construct even smarter systems. While OpenAI insists safety remains a priority, they are handling it differently now. Several safety staffers have already left the company, and some employees are openly worried OpenAI isn't being careful enough as its models become more powerful. Source: Financial Times / Writer: Daniyal

@not_a_charmOpenAI shut down the Preparedness team that was specifically built to monitor catastrophic AI risks. That group spent its time assessing how bad things could get if AI got weaponized for cyberattacks or biological threats, and now that work is being handed off to other teams. The former lead has pivoted entirely to building AI that can improve itself and help construct even smarter systems. While OpenAI insists safety remains a priority, they are handling it differently now. Several safety staffers have already left the company, and some employees are openly worried OpenAI isn't being careful enough as its models become more powerful. Source: Financial Times / Writer: Daniyal