eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
In a move toward the prediction that AI agents will soon be able to perform human tasks reliably, OpenAI has launched an agentic capability dubbed deep research that can handle complex tasks by conducting “multi-step research” online. The deep research agent can accomplish “in tens of minutes what would take a human many hours,’’ the company claimed.
How does OpenAI’s deep research tool work?
Once ChatGPT is given a prompt, OpenAI’s deep research agent will find, analyze, and pull together hundreds of online resources to create a comprehensive report that the company claims is on par with that of a research analyst.
“What you get from deep research is a comprehensive, fully cited research paper, essentially something that an analyst or an expert in a field might produce to you,” OpenAI Chief Research Officer Mark Chen said during a demo of the new deep research tool.
The agent uses reasoning to search, interpret, and analyze vast amounts of online text, images, and PDFs. Deep research will pivot when applicable to react to information it encounters, OpenAI said.
The agent was trained on real-world tasks requiring a browser and Python, “using the same reinforcement learning methods behind OpenAI o1, our first reasoning model,’’ the company said.
Although o1 is capable when it comes to coding, math, and other technical domains, OpenAI said deep research builds on reasoning capabilities to solve many real-world work and everyday challenges that demand more context and information gathering from diverse online sources.
The company issued a caveat that deep research is still in the early stage of development and “can sometimes hallucinate facts in responses or make incorrect inferences,” though at a lower rate than other ChatGPT models.
Who is the market for OpenAI’s deep research tool?
Deep research is aimed at people who do intensive work in areas including finance, science, policy, and engineering. However, shoppers who want very personalized recommendations on purchases that typically require a lot of research, such as cars, appliances, and furniture, will also find it useful, the company stated.
OpenAI’s Pro customers who pay $200 a month are the first people to be able to use this deep research tool.