OpenAI has announced a significant expansion of its ChatGPT platform with the introduction of Deep Research, a powerful AI tool designed to facilitate in-depth, complex research. This new capability is specifically tailored for individuals who require thorough, precise, and reliable research in areas such as finance, science, policy, and engineering.

Deep Research is intended for instances where users need to consider information from multiple websites and sources, rather than simply seeking a quick answer or summary. According to OpenAI, this feature will be particularly useful for individuals making significant purchases, such as cars, appliances, and furniture, that typically require careful research.

To utilize ChatGPT Deep Research, users will select the "deep research" option in the composer, enter a query, and have the option to attach files or spreadsheets. The AI system will then take anywhere from 5 to 30 minutes to provide a response, with users receiving a notification upon completion. Currently, Deep Research outputs are text-only, but OpenAI plans to add embedded images, data visualizations, and other "analytic" outputs in the future.

One of the primary concerns surrounding AI-powered research tools is their potential for errors and inaccuracies. To mitigate this risk, OpenAI has implemented a special version of its o3 "reasoning" AI model, which was trained through reinforcement learning on real-world tasks requiring browser and Python tool use. This fine-tuned model enables Deep Research to perform multi-step research, backtrack and react to real-time information, generate graphs, and specifically cite hundreds of sources and passages.

In testing, the o3 model powering Deep Research achieved an accuracy of 26.6% on Humanity's Last Exam, an evaluation that includes over 3,000 expert-level questions in various academic fields. While this may seem like a relatively low accuracy rate, OpenAI notes that Humanity's Last Exam is designed to be more challenging than other benchmarks to stay ahead of model advancements.

Despite its capabilities, OpenAI acknowledges that ChatGPT Deep Research has limitations. The system may struggle to distinguish authoritative information from rumors, fail to convey uncertainty, and make formatting errors in reports and citations. However, the company's commitment to providing fully documented outputs with clear citations and a summary of the thinking process aims to facilitate easy reference and verification of the information.

The launch of ChatGPT Deep Research raises important questions about the impact of generative AI on students and individuals seeking information online. While this feature has the potential to provide more accurate and reliable results, it remains to be seen whether users will subject the output to real analysis and double-checking, or simply treat it as a more professional-looking text to copy-paste.

Notably, Google announced a similar AI feature with the same name less than two months ago, highlighting the growing competition in the AI research space. As the technology continues to evolve, it will be important to monitor its development and potential applications across various industries.

In conclusion, OpenAI's ChatGPT Deep Research represents a significant step forward in AI-powered research capabilities, offering users a powerful tool for conducting in-depth, complex research. While the system is not without its limitations, its potential to provide accurate and reliable results makes it an important development in the field of artificial intelligence.