OpenAI has unveiled a new feature for ChatGPT called “deep research,” designed to autonomously conduct multi-step research tasks on the internet.
This tool allows users to submit questions with additional context through text, images, PDFs, or spreadsheets, and the system responds within 5 to 30 minutes, providing process summaries with citations.
Despite its capabilities, it can sometimes produce inaccuracies and confuse authoritative information with rumours. The feature is initially available on the web version of ChatGPT, with plans to release it for mobile and desktop apps later in February.
In a demonstration, OpenAI researchers showcased how the tool can synthesise web search data to recommend ski equipment for a holiday in Japan. OpenAI’s CEO, Sam Altman, is currently in Tokyo to meet with Japan’s Prime Minister, Shigeru Ishiba, and Masayoshi Son, the head of SoftBank Group. They are expected to discuss plans to enhance Japan’s AI infrastructure, including building AI data centres and power plants to support them.
Meanwhile, Chinese AI startup DeepSeek has emerged as a significant competitor in the AI field. DeepSeek’s R1 model has garnered attention due to its impressive capabilities, comparable to leading US models such as OpenAI’s, but achieved with a fraction of the budget.
Notably, DeepSeek utilises Huawei’s Ascend 910C chipsets for inference, which, while slower than Nvidia’s H800 chips, are better suited for AI tasks and offer cost advantages.
This development has raised concerns among US AI companies, as DeepSeek’s performance has sparked a wave of accusations that it has reverse-engineered the capabilities of leading US technology, such as the AI powering ChatGPT.
OpenAI has warned that Chinese companies are actively attempting to replicate its advanced AI models, prompting closer cooperation with US authorities.