With AI rapidly advancing, OpenAI, bolstered by Microsoft, is taking a democratic turn to keep its software in check. As announced by Reuters, the team at the AI lab is recruiting fresh talent to form a "Collective Alignment" team. This squad will focus on developing democratic processes to guide the governance of AI systems, aiming to address concerns such as bias and misinformation.
The move follows the conclusion of a grant program initiated in May 2023, which explored democratic process experiments. "As we continue to pursue our mission towards superintelligent models who potentially could be seen as integral parts of our society ... it's important to give people the opportunity to provide input directly," Tyna Eloundou, a research engineer and founding member of OpenAI's new team, told Reuters, showing commitment to involving the public voice in AI development.
With the 2024 U.S. election clock ticking, the specter of AI-generated "deepfake" content and the spread of misinformation is a valid concern. ChatGPT, a generative AI technology from OpenAI that can churn out convincingly authoritative text, has taken the public by storm since its 2022 launch. Yet, biases in such AI systems—which stem from the data fed into them—persist, raising alarms over the potential for racist or sexist output.
In response, OpenAI hints at a possible partnership with Worldcoin, a cryptocurrency endeavor by OpenAI CEO Sam Altman, to distinguish humans from AI bots in the voting framework. Giving the details, Teddy Lee, a product manager on the new team, mentioned a potential integration with Worldcoin, though they haven't set concrete plans for such an integration yet.
Meanwhile, the "Democratic Inputs to AI" grant program update details the journey and achievements of their investment in ten global teams. These groups strove to shape AI governance with groundbreaking democratic technology. Reaching across the digital divide and finding consensus amongst polarized viewpoints stood out as notable challenges, with a consistent realization that public opinion is a shifting sandcastle.
To address these challenges, OpenAI plans to refine and deploy systems that can gather and incorporate public input into the behavior of its AI models. "We are recruiting exceptional research engineers from diverse technical backgrounds to help build this work with us," the company said, eager to advance society's role in directing the ethics and actions of AI technology.
.jpg){kind=link}