The technology company OpenAI, known for creating the popular ChatGPT, has decided to eliminate the department responsible for analyzing future risks of Artificial Intelligence (AI) just one year after its creation, as reported by specialized U.S. media outlets on Friday.
Internal Reorganization
The department, known as Superalignment, was supposed to receive up to 20% of OpenAI’s computing power over the next four years to research and evaluate the potential dangers that could arise with advances in AI. However, it has been reported that employees from this division are being reassigned to other areas within the company.
Key Departures
The news of the department’s suppression comes shortly after Ilya Sutskever, co-founder of OpenAI, and Jan Leike, co-director of Superalignment, announced their departure from the company. Both argued that OpenAI should focus more on the security, oversight, readiness, and social impact of its technologies.
“Building machines smarter than humans is an inherently dangerous endeavor. OpenAI is taking on a huge responsibility on behalf of all humanity. But in recent years, security culture and processes have taken a back seat to innovative products,” Leike wrote on the X social network after his departure.
Changes in Leadership
In November, OpenAI’s board fired co-founder and CEO Sam Altman, citing alleged dishonesty, which caused significant uncertainty among investors and employees. However, a week later, Altman returned to the company, and board members who had voted for his dismissal, such as Sutskever himself, were removed from their positions on the board, although Sutskever remained in the company in another role.
Innovation vs. Security
Reports from media outlets like The Wall Street Journal pointed to an internal struggle between Altman and Sutskever. Sutskever, who was more focused on the impact of AI on humanity, clashed with Altman, whose primary focus was to continue with innovation.
New Releases
This week, OpenAI launched a new AI model, GPT-4o, along with a desktop version of ChatGPT and an updated interface, with the aim of facilitating user communication via video with their conversational bot.
The decision to eliminate the risk analysis department raises questions about OpenAI’s future approach to security and risk management associated with artificial intelligence, at a time when technology is advancing rapidly, and its impact on society is becoming increasingly significant.
Source: Noticias Inteligencia Artificial