Ms Martin, GenAI opens up a wide range of applications for businesses. Where do we stand today?
Andrea Martin: Depending on the study, between just over a third and half of companies in Germany use artificial intelligence. This is a significant increase compared to previous years. We see numerous applications that we at IBM also use ourselves as a so-called “client zero”:
Generative AI is used, for example, to generate, analy´se and test code. However, as I recently read, it’s important to note: “The idea that generative AI turns everyone into a developer is just as much a misconception as the idea that a camera turns everyone into a photographer.”
In human resources (HR), generative AI can help optimise processes, conduct conversations and handle routine tasks – with automation running in the background.
In customer support, generative AI is used to answer customer enquiries via chatbots and provide context-relevant information – in natural language, anywhere and around the clock.
In marketing, the creation of text, images and videos can be accelerated using generative AI by creating templates that staff can review and modify.
In addition, generative AI can of course also be used to prepare for consultations, e.g. between doctors and patients, by summarising findings, providing basic information on the clinical picture, or ‘translating’ technical terms into language that is generally understandable. There are numerous other potential applications, but that would go beyond the scope of this article.
Where are the opportunities and risks?
Andrea Martin: Generative AI has the potential to deliver significant productivity gains for both individuals and businesses. This can be achieved by optimising processes, supporting employees with routine tasks and fostering their creativity, as well as providing targeted support for task automation. Furthermore, generative AI can help improve a company’s products and services and enhance customer interactions.
One risk associated with the use of generative AI can arise, amongst other things, from a lack of transparency regarding the training data or the underlying and ‘hard-coded’ values. However, it should be possible to take these into account – particularly in a business context. In my view, therefore, sensibly implemented regulatory requirements regarding transparency, fairness and explainability are appropriate in order to avoid or mitigate risks.
Another fundamental risk associated with the use of generative AI is a lack of understanding of how the technology works: users of AI should at least be familiar with the basics of generative AI, for example that generative AI is based on statistical methods and probability theory and can, among other reasons, ‘hallucinate’. Only if I know this can I use generative AI responsibly.
How can we ensure responsible use?
Andrea Martin: There are many measures that can be taken to ensure the responsible use of generative AI. For me, there are three key points:
Firstly: Comprehensive AI governance is key to the responsible use of artificial intelligence, including generative AI.
Here are a few points to consider: The aim of AI governance is to ensure transparency regarding the objectives, benefits, data, etc. of an AI application throughout its entire lifecycle – from development and training through to implementation and operation. This also means monitoring compliance with the AI application’s objectives, as well as the accuracy of its responses. Furthermore, risks such as discrimination, manipulation of data and results, or breaches of data protection must be prevented or mitigated.
Effective AI governance also involves defining principles, establishing methods and processes, defining roles and bodies, and selecting technologies that support AI governance comprehensively or in specific areas.
Secondly: As mentioned above, users should have at least a basic understanding of how the technology works. This means that companies using or intending to use generative AI applications should invest in further training for their staff.
Thirdly: Companies should stay informed so that they can select reliable generative AI applications and avoid tools that raise concerns regarding ethical guidelines.
The discussion paper ‘Generative KI verantwortungsvoll einsetzen’ (in German) is available for download here.
The interview is released for editorial use (provided the source is cited © Plattform Lernende Systeme).