Advanced AI super-agents planned by OpenAI
Controversy brews over OpenAI’s use of a funded benchmark to highlight model capabilities.
OpenAI plans to introduce AI ‘super-agents’ designed to handle complex tasks at an expert level, according to a report by Axios. These advanced systems aim to perform intricate, goal-oriented tasks, far surpassing current AI chatbot capabilities. The announcement is expected within weeks, sparking widespread interest and scepticism alike.
CEO Sam Altman’s recent engagements in Washington DC, including a scheduled closed-door meeting with US officials, have intensified speculation. Social media rumours suggested a breakthrough in artificial general intelligence (AGI), prompting Altman to clarify that OpenAI has not developed AGI nor plans to deploy it soon. Despite this, the proposed super-agents are projected to be transformative, with potential applications ranging from software creation to business operations.
Critics argue the claims may be overhyped. Notable figures like computer scientist Gary Marcus dismissed the feasibility of achieving such advancements in the near term. Concerns about reliability and persistent issues like information hallucination remain significant barriers to broader adoption.
Controversy also surrounds OpenAI’s flagship AI model, o3, and its reliance on a benchmark test developed by Epoch AI, a group funded by OpenAI. The FrontierMath test, intended to measure mathematical prowess, has faced scrutiny over its role in showcasing the model’s capabilities.
About author
DW Team
The GIP Digital Watch Observatory team, consisting of over 30 digital policy experts from around the world, excels in the fields of research and analysis on digital policy issues. The team is backed by the creative prowess of Creative Lab Diplo and the technical expertise of the Diplo tech team.