ChatGPT could already replace 44 jobs according to an OpenAI study — but caution is still advised

Artificial intelligence continues to push the boundaries of what's possible. According to a new internal study published by OpenAI, the renowned creator of ChatGPT, AI models are already capable of partially or fully performing the tasks of 44 human jobs. This figure is both thought-provoking and worrying.

But behind the promises of efficiency and productivity, there are still areas of uncertainty: reliability, ethics, and long-term social consequences.

A new measure from OpenAI: the GDPval

To concretely assess the impact of artificial intelligence models, OpenAI has developed a tool called GDPval (for Generative Data Productivity Value).

This assessment helps to measure the extent to which models like ChatGPT can already replace certain human professions.

In an official blog post, Sam Altman's team clarifies:

“People often speculate about the overall impact of AI on society, but the clearest way to understand its potential is to examine what the models can already do.”

In other words, OpenAI aims to ground the debate in measurable facts, rather than assumptions or irrational fears. The stated goal is to provide a framework for reflection based on tangible evidence, particularly for the workplace and economic decision-makers.

44 professions already impacted by ChatGPT and its related technologies

Which professions are most threatened by the arrival of artificial intelligence?

The results of the GDPval study show that 44 jobs could already be automated or profoundly transformed thanks to AI.

Among them are administrative professions as well as technical or relational jobs:

Real estate agents
Social workers
Industrial Engineers
Software developers
Lawyers
Nurses
Customer service representatives
Pharmacists
Financial advisors
Private detectives

OpenAI cites concrete examples: an AI capable of evaluating images of skin lesions to assist a nurse, or the automatic generation of sales brochures for a real estate agent.

These capabilities illustrate how models can accelerate repetitive tasks, while reducing production costs and lead times.

ChatGPT: Between a support tool and a social threat

Officially, OpenAI insists on a reassuring message:

The goal is not to completely replace humans, but to support them in their daily work.

However, critics point out that this rhetoric masks a much broader ambition: to reduce labor costs and increase profitability on a large scale.

Behind the cautious rhetoric, Sam Altman's project – founder of OpenAI – seems clear: to make ChatGPT an essential player in global productivity.

The leader did not hesitate to claim that GPT-5 possesses “doctoral level intelligence”, before correcting himself in the face of initial tests deemed exaggerated.

A typical communication strategy for OpenAI, often criticized for its aggressive marketing.

Results to be interpreted with caution

While OpenAI's figures are generating discussion, many experts remain skeptical.

According to them, the GDPval assessment method still lacks transparency and is based on very theoretical scenarios.

“Most jobs cannot be reduced to a simple list of tasks that can be coded,” OpenAI itself points out in its report.

This is where the problem lies: human professions have an emotional, social and intuitive dimension that algorithms still struggle to reproduce.

Social workers, for example, do not just evaluate data; they listen, advise, and support.

Lawyers, for their part, must interpret the nuances of language and context, far beyond simply drafting a legal text.

AI still suffers from hallucinations and errors.

The biggest obstacle remains that of hallucinations, those errors in reasoning or facts produced by generative AI models.

ChatGPT, like other systems, can invent completely false quotes, sources, or conclusions with a confident and credible tone.

This instability makes its use risky in sensitive areas: medicine, law, finance, education…

Users often spend more time checking and correcting information than actually benefiting from a real productivity gain.

Several companies that have tried to partially replace their human teams with AI have faced disastrous consequences: legal errors, inconsistent documents, loss of customer trust.

Automation at all costs: a risk to social equality

Beyond technical performance, the question of social impact remains central.

If tools like ChatGPT become capable of performing the tasks of a real estate agent, an administrative assistant, or a writer, what will happen to the millions of people doing those jobs?

Some experts speak of a transition towards new hybrid roles, where humans would become “supervisors” of AI.

But in practice, low-skilled jobs are likely to be automated without real compensation.

The idea of a universal basic income or a tax on automation is therefore resurfacing, in the face of a technological revolution whose effects remain uncertain.

OpenAI: Between Ambition and Responsibility

Aware of the explosive nature of this debate, OpenAI is now communicating with more caution.

The San Francisco-based firm refers to a “human-machine collaboration” rather than a total substitution.

She points out that AI must remain tools under human supervision, especially for decisions with a strong ethical or emotional impact.

But in reality, the pace of development is accelerating.

With the release of GPT-5, announced as “the most advanced model ever created”, OpenAI confirms its ambition:

extend the uses of ChatGPT to medicine, education, law, artistic creation and scientific research.

This expansion raises another question:

Should the use of generative AI be more strictly regulated before it redefines the job market?

Towards a new definition of human work?

Looking further, the OpenAI report raises a fundamental question:

What truly defines a human profession?

Is it technical skill, intuition, creativity, or the ability to empathize?

Artificial intelligence forces us to rethink these concepts.

Perhaps in the long run, jobs will not disappear, but will evolve towards a form of cooperation between man and machine.

The real question will not be whether ChatGPT can replace humans, but how we wish to use this technological power — and within what ethical framework.

In conclusion

Yes, ChatGPT can already replace certain specific tasks in 44 professions according to OpenAI.

But no, that doesn't mean that humans have become obsolete.

The GDPval study highlights above all the duality of artificial intelligence: capable of helping, but still fragile, prone to errors and lacking moral intuition.

The future of work will not be about replacements, but about intelligent adaptations.

And the real question we need to ask ourselves is not “how far will ChatGPT go?” but “how far do we want to let it go?”

The 10 jobs most threatened (or transformed) by ChatGPT and AI

OpenAI's GDPval study and labor market expert analyses agree: some professions are more exposed than others to automation by generative AI.

Here are the Top 10 jobs most likely to change significantly by 2030:

Rank	Job	Expected impact	Tasks involved
1️⃣	Administrative Assistant	Very strong	Writing emails, managing schedules, creating reports
2️⃣	Web editor or journalist	Pupil	Content production, SEO, automated summaries
3️⃣	Real estate agent	Strong	Creating advertisements, drafting contracts, calculating estimates
4️⃣	Financial advisor	Strong	Portfolio analysis, investment simulation
5️⃣	Software developer	Medium to high	Code generation, testing, documentation
6️⃣	Lawyer or legal professional	AVERAGE	Contract drafting, legal research, case law analysis
7️⃣	Nurse and pharmacist	Moderate	Preliminary diagnosis, medical image analysis
8️⃣	Community manager	Strong	Creating posts, moderating, scheduling
9️⃣	Translator	Very strong	Machine translations and assisted revisions
🔟	Customer Service Representative	Very strong	Automated responses, request tracking, AI chatbots

💡 Good to know: these jobs won't disappear overnight. They will evolve towards management, supervisory, or value-added roles that AI cannot yet replicate: empathy, strategy, moral judgment.

🔸 How to prepare for work

with

AI, not against it

Faced with the rapid rise of tools like ChatGPT, adaptation becomes key.

Here are some tips to anticipate the change:

Develop your AI skills: learning to use tools like ChatGPT, Midjourney or Claude is becoming as essential as knowing how to use Word or Excel.
Strengthen your critical thinking skills: AI can make mistakes. Knowing how to validate, cross-check, and correct the information generated will make you an "augmented" user.
Showcase your creativity and emotional intelligence: two qualities that algorithms cannot master.
Train yourself in digital ethics and data protection: these topics will be at the heart of the jobs of the future.
Use AI as a partner, not a substitute: delegate repetitive tasks to it, keep the strategy and vision for yourself.

🔸 In summary

Artificial intelligence does not kill work, it reshapes it.

The jobs of tomorrow will be half-human, half-algorithmic, where value will reside less in execution than in understanding, creativity and responsibility.

OpenAI, with ChatGPT and its GDPval tool, is paving the way to a future where human-machine collaboration will be the new engine of global productivity — provided we do not forget that progress is only worthwhile if it benefits everyone.