If all you have is a hammer, the world seems full of nails. Today, that hammer is called generative AI, and it is shaping the way we think about artificial intelligence. Companies are racing to identify use cases for generative AI, also known as GenAI. The chemical and pharmaceutical company Bayer, for example, has identified more than 700 cases. But this also includes analysing Excel results and creating a draft text in Word.
Usecase Inflation
If all you have is a hammer, the world seems full of nails. Today, that hammer is called generative AI, and it’s changing the way we think about artificial intelligence. Companies are in the process of identifying use cases for generative AI, also known as GenAI. The chemical and pharmaceutical company Bayer, for example, has already identified more than 700 such cases. This includes analysing Excel results and creating a draft text in Word.
Companies are hoping to attract more attention with the AI label and want to position themselves on the topic of the future. This is working quite well for the manufacturers of the tools, as can be seen on the capital market. The Economist has identified a chatbot premium. Since the launch of ChatGPT, the tech stocks in the S&P have outperformed the other S&P stocks by almost 40 %.
Maslow’s Hammer
As always in a gold rush, the first to benefit are the manufacturers of the tools. This time it seems to be no different. If you want to mine AI gold, you need GenAI applications. We are seeing a run on these beautiful new tools. GenAI applications promise to turn data into gold.
But the gold rush threatens to blind us. As psychologist Adam Maslow put it in 1966: “I think it’s tempting, when the only tool you have is a hammer, to treat everything as if it were a nail”.
Today, GenAI applications are used for many things that look like a nail. They are useful in areas such as programming, customer service and text analysis. However, they reach their limits when it comes to tasks that require a clear result.
Not every problem requires a hammer
This is because GenAI models generate answers based on a variety of factors, including the use of the trained dataset, user input and internal probability models. As a result, these models will provide different answers to the same query. However, this variability is essentially their strength, as it allows for a wider range of creative or contextual answers. It also means that GenAI is not suitable for applications that require consistency and predictability.
GenAI’s strength is not precision and consistency
Not every problem needs a hammer. If you are looking for clear and repeatable answers, other AI methods or algorithmic approaches are more appropriate. For example, rule-based systems, decision-tree algorithms and structured database queries are deterministic: they produce consistent results under the same conditions and inputs. These methods are ideal for applications where precision and consistency are critical, such as accounting, legal analysis or safety-critical systems.
Despite all the euphoria about the new GenAI tool, it’s worth taking a look in the toolbox from time to time and asking yourself whether you need a hammer for everything. Or better still, close the toolbox and take a moment to think about the problem that needs solving.