This article invites you to explore what ChatGPT is, how it works, and what it means for the future of workers.
*the DeLorean time machine is a time travel vehicle constructed from a retrofitted DMC DeLorean, in the Back to the Future franchise.
Generative artificial intelligence (AI) – embodied here by ChatGPT – is at the heart of a technological revolution that is disrupting the world of work, and white-collar jobs in particular. This article invites you to explore what ChatGPT is, how it works, and what it means for the future of workers. It is a journey without a DeLorean, but one that propels us into a fascinating and complex future, where technology is redefining what it means to work.
ChatGPT is an AI tool created by OpenAI, an organization founded in 2015 with the aim of developing safe artificial intelligences that benefit humanity. The term “GPT” stands for Generative Pre-trained Transformer. In simple terms, it is a statistical model that has learned to understand and produce natural language by analyzing thousands of billions of words taken from the Internet.
OpenAI has developed a series of versions of GPT, each iteration proving more sophisticated than the last. ChatGPT, in its current form, is based on GPT-4o, a model that is capable of answering complex questions, generating content and even solving problems creatively (and I mean creatively….). What sets it apart from other automatic natural language processing tools is its ability to create answers that sound human, using a large corpus of texts to learn grammatical structures and concepts.
The heart of ChatGPT is based on neural networks, a technology inspired by the human brain. Imagine a complex network of artificial neurons, each connected to other neurons, working together to transform simple sentences into complex ideas. Here is how the process works, in simple terms:
Training: The model is pre-trained on a large number of texts. During this training, ChatGPT does not acquire direct knowledge like a human, but learns regularities, patterns and relationships between words and sentences.
Prediction: When ChatGPT is asked a question, it predicts the most likely answer based on the data available. In practice, it constructs its answer word by word, each word depending on the context given by the previous ones.
Refinement: One of the specific features of ChatGPT is that it has been refined by a method called Rétroaction Apprise par Renforcement (or RLHF, Reinforcement Learning from Human Feedback). This means that humans have evaluated ChatGPT’s responses, helping it to improve its accuracy and relevance.
ChatGPT essentially works by being able to understand the context of questions and produce coherent answers. It does not “understand” language like a human, but its learning on a large amount of data enables it to efficiently manipulate information to generate relevant answers.
The rise of ChatGPT and other generative AI technologies raises alarming questions about the future of white-collar jobs – those who work primarily in offices, use their intellect to complete tasks, and are not involved in physical manual labor. These jobs are particularly prevalent in the administrative, financial, legal and content creation sectors. Here are a few concrete examples of possible impacts:
Administrative assistants and office workers
Administrative assistants who organize meetings, answer emails and prepare documents are already under threat from AI. Tools like ChatGPT can write formal emails, summarize meeting notes, and even manage calendars, reducing the need for a human presence for these tasks.
Imagine an employer who needs to cut costs. Rather than hiring a new assistant, it could simply subscribe to a service that uses AI to perform these tasks more affordably. This directly shifts workers into unemployment or forces them to switch to other types of employment.
A study by McKinsey & Company showed that automating administrative tasks with AI can increase productivity by 20-30%. For example, one company that integrated AI tools for email and calendar management saw a 25% reduction in the time spent on repetitive tasks, enabling employees to focus more on high value-added activities.
Journalists and content creators
Another area where ChatGPT is making an impact is in journalism and content creation. Blog posts, news reports or simple financial analyses can be produced by AI at a speed and cost that defy human competition.
Some major media outlets have already started experimenting with AI-written articles, mainly to cover basic news where human creativity adds no particular value. This poses a risk not only to journalistic jobs, but also to the quality of information, as AIs can reproduce biases present in their training data.
Financial sector: analysts by the numbers
Financial analysts often use data to produce reports or advise on investment strategies. With AIs capable of analyzing huge volumes of data, detecting trends, and providing advice, these roles are also under threat. For example, AI-based platforms can already produce financial valuations or portfolio recommendations, reducing the need for human experts in simple cases.
Faced with these revolutions, the question for trade unions becomes: how to protect workers’ rights and prepare employees for this ever-changing future?
Upgrading the Skills and Training
One of the key responses to the impact of AI on jobs is requalification. Unions must collaborate with employers and educational institutions to ensure that displaced workers can acquire new skills in sectors that are not threatened by automation.
Negotiating agreements on further training should become a priority, with an emphasis on complementary skills that AI cannot easily replace, such as creativity, human management, and tasks that require empathy.
Regulating the use of AI
It is also necessary to regulate the use of AI in the workplace. Unions need to be involved in developing policies that protect workers from unfair dismissal due to automation. For example, by negotiating longer transition periods before AIs can replace human jobs, or by establishing an ethical framework for the use of these technologies.
Maintaining human intervention
Unions could also argue for a hybrid approach, where AIs like ChatGPT complement the work of humans rather than replace them. This could mean that, for example, a journalist works with AI to generate the most basic parts of a report, leaving the human to provide critical analysis and added value.
Imagine a world where only human leaders remain, surrounded by machines performing all other tasks. It is a future that may sound like something out of a science-fiction novel, but it has made possible by rapid advances in artificial intelligence. In such a scenario, machines would perform all productive work, while managers would make strategic decisions, retaining economic and political control.
This world could have advantages, such as almost unlimited productivity, the disappearance of tedious or repetitive tasks, and optimal efficiency. However, it is important to consider the social and human consequences of such a model. Without jobs, how would the general population find meaning in their lives or a means of sustaining themselves? Such a world would also be profoundly unequal, with a concentration of power and wealth in the hands of a ruling minority. The rest of the population could find themselves at the mercy of decisions taken by an elite, with no possibility of influence or contribution.
This hypothetical future raises questions about the value of work and the role of human beings in society. Work is not only a source of income, but also a means of self-fulfillment, of contributing to the community, and of building social bonds. If we delegate everything to machines, we risk losing these essential aspects of the human condition. For trade unions, the challenge is to ensure that technological advances benefit everyone, not just a minority, while preserving the dignity and usefulness of everyone in society.
Prepare for a trip without a DeLorean
Generative AI, pioneered by ChatGPT, is a transformative force redefining the future of work, especially for white-collar workers. While the prospects may seem daunting, they also offer opportunities to redefine the role of workers and adapt the organization of work to be more resilient and equitable.
For trade unionists, the challenge is twofold: to ensure that workers can enjoy the benefits of AI while preparing for the disruption it will inevitably bring. The key lies in education, regulation and collaboration between all stakeholders. This journey into the future, without DeLorean but with innovative technologies, is an invitation to be alert, adapt, and collectively determined to build a future where humans remain at the centre of the new value it generates.
In a forthcoming article, we will discuss education and medicine, which will be the two fields most impacted by AI, for better and for worse.
Rudy Druine is a long-standing USB member and former Head of Operations and IT Resource Management at the European Commission; he is currently Professor of Computer Science at the Institut d’Enseignement Technique Commercial de Promotion Sociale de la Province du Hainaut.