OpenAI Aims to Build AI That Researches Itself

Leading artificial intelligence (AI) company OpenAI has set a top priority to enable AI to conduct scientific research on its own over the next few years. This is because they see potential for AI to solve complex scientific problems and accelerate new discoveries. To secure technological leadership, OpenAI is dedicating all its efforts to building a fully automated AI researcher.

OpenAI announced that by September, it plans to build an 'autonomous AI research intern' capable of handling specific research problems independently. Following this, by 2028, it plans to unveil a multi-agent-based, fully automated research system. This system is expected to address problems that are too large or complex for humans to handle, such as in mathematics, physics, life sciences, business, and policy dilemmas.

OpenAI stated that it is concentrating all its capabilities on building a 'fully automated AI researcher' that can conduct research independently. This system is expected to possess the ability to process not only text and code but also complex problems written as whiteboard scribbles. Jakub Pachocki, OpenAI's Chief Scientist, mentioned that they are approaching the point where they will have models capable of working indefinitely and consistently, much like humans do. The research team, including Chief Scientist Pachocki and Chief Research Officer Mark Chen, is conducting research with the goal of automating entire research labs within data centers.

Inference models currently underpin all major chatbots and agent-based systems. OpenAI's technical staff are utilizing these systems in their work, viewing them as the foundation for the fully automated researchers that will emerge in the future. OpenAI is engaged in fierce competition with rivals such as Anthropic and Google DeepMind.