With the success of ChatGPT, more and more companies are posting positions for so-called prompt engineers. We are looking for AI whisperers who make digital user interfaces even more intelligent. IT companies are no longer the only ones who can afford this.
There is good money to be made in the tech industry. The is no longer a secret. Last but not least the Hype about artificial intelligence (AI) has also fueled this trend. The success of ChatGPT is now focusing on a job that very few had on the screen so far – and for which hefty salaries are paid.
So-called prompt engineers train AI chatbots and develop commands that can be used to achieve even better results. The AI whisperers use speech recognition technologies to process and interpret user input (so-called prompts) and generate corresponding responses or actions. “Prompt engineers thus enable easy and improved access to language models. They work as a kind of translator between users and AI,” says Alexander Brem, Professor of Entrepreneurship in Technology and Digitization at the University of Stuttgart, ntv.de.
And that costs companies a lot. The Google startup Anthropic, for example, has just advertised a position for a prompt engineer with an annual salary of 335,000 US dollars, according to the finance portal Bloomberg. Applicants should be familiar with the common language models (LLMs) and have basic programming skills. Californian firm Klarity is offering $230,000 for a similar position.
Use no longer only in IT companies
The top-paying jobs often go to people with PhDs in machine learning or those who have already founded AI companies. Apparently, applicants without a technical degree also have a good chance of getting such a lucrative job. According to Brem, graduates of non-technical courses – such as linguistics, linguistics, communication sciences or psychology – can be particularly important for input processing. “Finally, it must be ensured that the AI model can reliably process a large number of user inputs into the respective Conversational User Interface (CUI). In this respect, you need well-educated people who can understand things very well or who can contribute technical expertise through experience ” says Brem.
Prompt engineers are not only in increasing demand, they are no longer only used in IT companies. Financial firms, law firms, insurance companies, editorial offices and advertising agencies are now also looking for them. They are needed wherever customer or user interaction is important and can be optimized through interactive CUIs. “Depending on the use case, the existing ‘generic’ language models for communication have to be adapted and optimized before they can interact with users, improve processes, collect feedback or answer questions,” says Sophia Hess, a PhD student at Brem. In the foreground The focus is always on optimizing the user experience. “If prompt engineers do a good job, it can increase AI and user adoption tremendously.”
In order to meet the demand, the PromptBase platform was launched last year. The online marketplace places prompt engineers and sells successful command lines. The job may still only be known to a few. But experts are already making comparisons to blockchain, non-fungible tokens and cryptocurrencies due to the rapid growth of the industry. While they were still experiencing a job boom in 2021 and attracted applicants with high salaries and bonuses, many talents have already left the crypto world.
AI learns, but prompt engineers are not superfluous
According to Brem, how long the run on prompt engineers will last depends, among other things, on the development of new technologies, the speed with which CUIs are disseminated and the demand for customer-oriented user interfaces. However, as long as companies continue to improve their interaction with customers and users and want to automate processes, prompt engineers will also be in demand. Brem expects her role to evolve over time as technology advances and new areas of application are discovered.
Although the AI learns with every request, that doesn’t mean that prompt engineers are working to make their own job redundant. “Since the queries in current models cannot influence the underlying AI model, prompt engineers do not have to worry about their professional future,” says Brem. There are two main reasons for this.
On the one hand, the processing of inputs that contain subtexts such as irony or sarcasm is a major challenge. On the other hand, the needs and demands of users are constantly evolving as technology advances. “In the near future, the role of prompt engineers is likely to become more complex and demanding as technology and its applications continue to advance,” says Brem.