OpenAI releases ‘Deep Research’ amid growing popularity of China’s DeepSeek

SINGAPORE: OpenAI has launched Deep Research, a new artificial intelligence (AI) tool designed to streamline online research, as competition from Chinese AI developers like DeepSeek intensifies. On Feb 2, the company said that the service, available to select paying customers through OpenAI’s ChatGPT chatbot, “find, analyze, and synthesize hundreds of online sources to create a comprehensive report at the level of a research analyst” in minutes.

OpenAI claims Deep Research is meant to complete tasks in “tens of minutes” of what would have taken a human many hours.

In its statement, OpenAI said Deep Research was designed for professionals in fields like finance, science, policy, and engineering who need precise and reliable research. The company added that the tool helps save “valuable time” by allowing users to expedite complex, time-intensive web research with just one query.

According to Bloomberg, Deep Research is OpenAI’s second AI agent this year. In January, the company launched Operator, which helps users book flights, plan grocery orders, and make purchases.

Both tools are currently limited to ChatGPT Pro users, who pay US$200 (about S$273) per month.

OpenAI-backer Microsoft, Anthropic, and other startups have also introduced similar technology, aiming to boost productivity by saving users time for personal and professional tasks.

OpenAI CEO Sam Altman has called AI agents the “next giant breakthrough.” The competition has intensified as Chinese AI companies, particularly DeepSeek, made rapid progress, raising concerns that they are catching up with leading American developers.

OpenAI warned that Deep Research is still in its early stages and has limitations. “It can sometimes hallucinate facts in responses or make incorrect inferences, though at a notably lower rate than existing ChatGPT models, according to internal evaluations,” the company said.

The company added that it might struggle to distinguish authoritative information from rumours, often failing to express uncertainty accurately, and could have minor formatting errors in reports and citations.

OpenAI said they expect all these issues to quickly improve with more usage and time.

Bloomberg reported that initially, users will be limited to just 100 queries per month. The company also plans to eventually extend access to its Plus, Team, and Enterprise subscribers, though it hasn’t provided a timeline. /TISG

Read also: Will DeepSeek shake the AI market further? The company claims its new Janus-Pro image generator outperforms DALL-E and Stable Diffusion