How ChatGPT & Co. Will Change Our Work: Four Hypotheses About Working with (Generative) Artificial Intelligence

Have you found yourself asking which of your daily tasks could be done by artificial intelligence (AI) – whether at home or at the workplace? I can think of a few things. For example, I could have used ChatGPT to write this blog post for me (but I didn’t).

While artificial intelligence sifts through, structures and analyzes vast amounts of (unstructured) data, generative AI can even draw conclusions and create new content, based on the data that’s been used to train it. Since the AI-based chatbot ChatGPT was released to the general public at the end of November 2022, the topic of (generative) AI has been on everyone’s lips. After just five days, one million people had registered for ChatGPT; after two months, the number of registered users had reached 100 million. By comparison, Spotify took over four and a half years to cross the 100 million user threshold; Instagram took two and a half years and even TikTok took nine months. Only Threads, the new microblogging service, grew even faster than ChatGPT in its initial weeks.

Many people are therefore saying that artificial intelligence is having an “iPhone moment”, which is when, practically overnight, an already well-known technology is used by the masses on an everyday basis because a new, relatively simple product makes it easy for anyone to operate.

How will the increasing use of ChatGPT and other (generative) AI applications impact our working environment? And how will it change the nature of our work? Of course, at Merck KGaA, Darmstadt, Germany, we are also faced with these questions. I would like to summarize my insights in four hypotheses:

1.    Automation of work processes is yesterday’s news – today, AI supports well-trained knowledge workers

Using AI is not about automated production processes in factories, i.e. robots, taking over people’s jobs. It is rather a matter of AI gathering and processing unstructured data and information. As such, it also has the potential to support well-trained knowledge workers.

A study by Goldman Sachs concludes that a good two thirds of occupations could be partially automated by AI in the future; only around 7% of jobs would be completely replaced by AI. Scientists at OpenAI – the start-up that developed ChatGPT – and the University of Pennsylvania assume that AI language models will change most jobs in some way or another. Around 80% of employees in the United States work in jobs in which at least one task could be carried out faster by generative AI. The scientists expect around 19% of employees to see at least half of their work tasks affected. Both studies conclude that AI applications will increasingly be used in offices and in administration but will see less use in skilled trades and in the operation of plants and machines where many processes are already automated.

2.    Job profiles are changing – from “creator” to “reviewer”

As AI supports our work to an ever-greater extent, the challenges we face will change. After all, some of the tasks that we’ve always had to do ourselves will be taken over by AI in the future, at least to a certain point. In the future, we will work less as “creators” of content such as letters, inventory lists and production plans because the AI can do this much quicker. Instead, we will increasingly take on a “reviewer” role and check whether the content created by the AI is correct and logical.

3.    Enabling instead of prohibiting – the management is paving the way

Since the AI can relieve us of many time-consuming activities, we need to learn to delegate tasks to it and use it as productively as possible. In other words, saying “no” to AI is not an option; quite the contrary, everyone should include these applications in their portfolio of work tools wherever they are useful and helpful.

Some companies are still trying to prohibit the use of ChatGPT out of fear that internal or even confidential information could be shared in a public network. This is understandable but also too short-sighted. Instead of blocking these innovations, I believe the more promising path forward is to enable employees to use them productively within the company in accordance with the rules. This is the path we have chosen at Merck KGaA, Darmstadt, Germany, having recently installed our own version of ChatGPT. Incidentally, we’re one of the first companies to do this.

4.    ChatGPT is just the beginning – connecting it with other systems gives rise to unimaginable possibilities

Even if companies can’t get around ChatGPT, plug-ins give them a chance to freestyle. Using plug-ins – i.e., extensions that enable ChatGPT to access other programs – ChatGPT can obtain search engine results, create travel itineraries, compare prices, execute code, access documents, and much more. For example, connecting an internal chatbot, to an ERP system that manages the business processes of a company gives rise to completely new possibilities in manufacturing, logistics, supply chain, and accounting.

All of this is coming – the development towards artificial intelligence cannot be stopped. The dynamism and speed with which new AI applications are coming onto the market is massive and the expectations associated with it are great. AI will sharply increase the productivity of humans and processes and ultimately, the quality, innovative capacity, and delivery reliability of our products. We therefore see AI as a crucial instrument for maintaining a competitive edge.

Of course, increased use of (generative) AI is also associated with risks – especially of an ethical nature. For example, users looking at images and texts may no longer be able to distinguish between fact and fiction, or they might not know whether they’re interacting with a human or an AI – which can generally lead to a loss of trust in communication. At Merck KGaA, Darmstadt, Germany, we have therefore developed a Code of Digital Ethics, which provides our colleagues with clear principles on how they should handle ethical issues in a digital context. Moreover, these principles are integrated into critical processes, meaning the ethical issues are taken into consideration right from the start.

This just leaves one crucial question: How will AI impact the employment situation? At the beginning of this blog post, I quoted studies indicating that AI will change jobs rather than causing them to disappear completely. I am convinced that AI, in most cases, will not replace humans, but humans who can deal with AI will replace those who can’t or won’t.