JAKARTA - OpenAI's artificial intelligence Laboratory is forming a new team, "Collective Alignment," which aims to create a democratic process to regulate its artificial intelligence software to address biases and other factors. This was conveyed by the Microsoft-backed company via a blog post on Tuesday, January 16.
The "Collective Alignment" team is a continuation of the grant program to fund the democratic process experiments announced in May 2023, as disclosed by the San Francisco-based company. The program recently ended.
"As we continue to pursue our mission towards a potential superintelegen model that can be considered an integral part of our society... it is important to give people the opportunity to provide input directly," said Tyna Electroduu, a research engineer and founding member of OpenAI's new team., quoted by VOI from Reuters.
"To ensure, for example, that only humans can vote, OpenAI can partner with Worldcoin, a cryptocurrency project founded by OpenAI CEO Sam Altman, that provides a way to differentiate between humans and AI bots," said Teddy Lee, a product manager who is another member of the team. the two of them.
Lee stressed that this team has not made concrete plans to integrate Worldcoin.
SEE ALSO:
Since being launched by OpenAI at the end of 2022, ChatGPT's generative artificial intelligence technology, which is capable of producing very convincing writing from text stimuli, has caught the public's attention and has become one of the fastest growing apps of all time.
However, there are concerns about AI's ability to create "deepfake" images and other false information, especially as the US 2024 election campaign is heating up. Critics say that AI systems like ChatGPT have inherent biases as inputs are used to shape their views, and users have found examples of racial or sexist AI outputs.
OpenAI's new team is currently actively looking for a research engineer and research scientist, said Eloundou. The team will work closely with OpenAI's "Human Data" team, which builds infrastructure to gather human input on the company's AI model, and other research teams.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)