On Feb 2, OpenAI launched its Deep Research feature, which can synthesise content from across the web into a detailed report. And it does so in minutes. And it’s powered by a version of the company’s latest model, o3.
This feature is a powerful tool for professionals. It allows them to save hours of work by performing research autonomously. But can the underlying model of the technology replace them? Yes, suggests Deep Research.
Min Choi, an X user whose account is dedicated to sharing AI content, asked Deep Research to “make a list of 20 jobs where the OpenAI o3 reasoning model will replace humans.
Deep Research just dropped a wild list…
20 jobs that OpenAI o3 will replace humans. pic.twitter.com/rB45IUqheC
— Min Choi (@minchoi) February 4, 2025
After diving into 24 sources in seven minutes, Deep Research produced a table that includes job titles, explanations of why an AI is better than a human in that role, and the likelihood that the position will be replaced.
What jobs will AI replace, according to ChatGPT?
Just in time for tax season, the role of “tax preparer” tops the chart with a 98% probability of replacement, which ChatGPT considers “nearly certain automation.”
ChatGPT explains that AI would be better at this task because it can quickly process tax rules and calculations, making it faster than a human. To support its argument, ChatGPT points out that AI-driven tax software already exists.
The other professions mentioned in the list are, in order, the following:
- Data Entry Clerk
- Telemarketer
- An accountant
- Legal technician
- Appointment organiser
- Virtual Assistant
- Transcriber
- Proofreader
- Editor
- Customer Service Representative
- Email Marketing Specialist
- Content Marketing Specialist
- Social Media Manager
- Translator
- Technical Support Analyst
- Recruiter
- Market Research Analyst
- Travel agent
- Tutor
All of the jobs offered by ChatGPT have a common theme: the role relies primarily on a technical skill that AI can perform well autonomously. For example, AI can transcribe, translate, and proofread efficiently, making the need for humans to perform the technical skills of these jobs less compelling.
AI can also analyse large numbers of documents, draw conclusions, and even take action based on its own analysis. Therefore, roles that make this process a core part of their responsibilities, such as social media managers, recruiters, travel agents, market research analysts, appointment schedulers, accountants, and legal assistants, are considered more at risk.
Are you doomed?
Does this study mean that everyone working in these fields will lose their jobs?
Nah, my friend.
ChatGPT’s analysis only considers AI’s ability to effectively perform technical skills. However, soft skills, including communication, critical thinking, conflict resolution, leadership, time management, and interpersonal skills, are just as important for success. And AI cannot replicate these abilities.
Additionally, organisations are not yet ready to adopt AI at scale. For example, a recent report from Accenture shows that many barriers still prevent business leaders from implementing AI at scale across organisations, citing lack of clarity on ROI, infrastructure limitations, and disconnect with employees.
How to try Deep Research yourself
If you want to ask a similar question to Deep Research or delve deeper into another topic, you’ll need to sign up for a $200 monthly subscription to ChatGPT Pro. While the price is steep, the subscription comes with other perks, such as unlimited access to ChatGPT and Sora, and access to Operator, its AI agent feature that can perform basic browser-based tasks like bookings.
There are also cheaper alternatives, such as Google’s Deep Research feature, which is available to all Gemini Advanced users via the Google One AI Premium subscription that costs $20 per month. You can also find Hugging Face’s recently released equivalent, also called Deep Research, which has a free demo.