Jakarta, Indonesia Sentinel — OpenAI has introduced Deep Research, a new artificial intelligence (AI) feature capable of conducting in-depth analysis within minutes, mimicking the expertise of human professionals. While the innovation promises to revolutionize research, it also raises concerns about AI potentially replacing human experts.
Deep Research is available as a feature in ChatGPT Pro, marketed as an advanced research assistant that can rival trained analysts. The AI capable to autonomously searches the web, gathers sources, and delivers structured reports.
The model has even scored 26.6% on Humanity’s Last Exam (HLE), a rigorous AI benchmark which outperforming many existing models. However, Deep Research is far from perfect and comes with significant limitations.
Accuracy Issues and Limitations
According to Science Alert, Deep Research struggles with key challenges such as missing critical details, difficulty accessing the latest information, and occasionally fabricating facts.
OpenAI has acknowledged these shortcomings found on Deep Research, stating that the tool “may sometimes hallucinate responses or draw incorrect conclusions,” though at a much lower rate than existing ChatGPT models, according to internal evaluations.
Can AI Replace Human Experts?
Unlike standard chatbots that provide quick responses, Deep Research follows a multi-step process to produce structured reports:
- User submits a request – This can range from market analysis to legal case summaries.
- AI clarifies the task – It may ask follow-up questions to refine the research scope.
- Autonomous web search – The AI scans hundreds of sources, including news articles, research papers, and online databases.
- Findings synthesis – The AI extracts key points, compiles a structured report, and cites sources.
- Final report delivery – Within five to 30 minutes, users receive a multi-page document that could resemble a PhD-level thesis.
While this appears to be an ideal tool for researchers and knowledge-based professionals, closer examination reveals major limitations.
Read Also:
OpenAI Launches ‘Deep Research’ Feature for ChatGPT, Complete Research in 5 Minutes
Key Shortcomings of Deep Research
- Lack of contextual understanding – AI can summarize and generate research but does not fully grasp what is truly important.
- Ignores recent developments – The AI has been found to overlook significant legal rulings and scientific breakthroughs.
- Fabrication of information – Like other AI models, Deep Research can generate inaccurate yet convincingly presented data.
- Inability to differentiate fact from fiction – The AI struggles to distinguish authoritative sources from unreliable ones.
The Future of AI in Research
Despite OpenAI claims that Deep Research can rival human analysts, AI still lacks the judgment, rigor, and expertise required to produce high-quality research outputs. While the tool offers a promising step forward in AI-assisted research, it remains far from replacing the nuanced insights provided by human professionals.
(Raidi/Agung)
