OpenAI and DeepMind employees publish letter criticizing AI companies

A group of experts, including current and former employees of companies operating in the field of artificial intelligence (AI), published a report. Open letter about the dangers of this technology. The text has a very critical tone towards companies that are currently developing tools for the sector.

13 people signed the letter titled “The right to warn about advanced artificial intelligence”. Of these, Mostly former employees OpenAIOwner of ChatGPT.

The other two participants (one currently works there and the other has already left) are from Google’s AI lab DeepMind, one of whom also worked for chatbot Claude Anthropic.

What does the open letter say about artificial intelligence?

Overall, the text is a critique of the way AI companies currently deal with this technology on several fronts, including transparency. After contextualization and general notes, the group asks companies in the region to agree to comply with four operating principles.

The authors say they are aware of the potential and benefits of artificial intelligence, as well as its risks. In fact, their biggest concern is the lack of openness from companies that share information only with their own employees (e.g. governments, independent organizations and society itself).

“AI companies have strong financial incentives to avoid effective oversight, and we do not believe that privatized corporate governance structures are sufficient to change this situation,” he says.

How are the activities of artificial intelligence companies regulated?

Principles advocated by the group include companies not entering into agreements that prohibit criticism of them by employees, having a reliable and anonymous process for warning and reporting risks, and supporting a culture of open criticism.

Additionally, those responsible for AI systems should allow employees to publicly report risks associated with the technologies, including sharing risk-related information even if confidential, after all previous safeguards have failed.

Recently, OpenAI created a Safety and Security committee to ‘oversee’ AI operations on its own. But even after receiving the actress’ rejection, it was she who used a voice similar to Scarlett Johansson’s in the chatbot.

So far, companies DeepMind, OpenAI and Anthropic have not commented on the open letter. You can read the entire text (in English) from this link.