AI applications are increasingly taking decisions out of people’s hands. Yet various incidents have already driven distrust of artificial intelligence very high. For the technology to gain broad acceptance, the developers of AI systems must restore trust. Only when we can be sure that AI applications are unbiased and handle our data with care can widespread adoption be achieved in the future. We explain how we can trust AI applications and what opportunities this opens up for companies.

AI applications can carry xenophobic or discriminatory views

Artificial intelligence applications have long been part of everyday life. Algorithms decide on loans, run application processes, detect diseases and identify suspects. Yet there are repeated cases in which AI applications discriminate against people for a wide variety of reasons. The problem lies in the data the algorithms learn from. If the data set is already biased, this manifests itself in the AI’s decisions. AI applications can incorporate xenophobic or discriminatory views from the programmers either consciously or unconsciously.

To prevent discrimination against users by AI systems, developers must have their programs reviewed by independent bodies. An in-house ethics committee or external advisors can serve this role. The same applies to the data used for machine learning. A careful review here is essential from the outset. Beyond the databases, the software developers who build the algorithms must also take ethical and moral principles into account.

Our children will expect a bias-free technology they can trust.

Bias-free AI: The key drivers

Today’s AI skepticism is a powerful driver pushing for greater sensitivity in the development of AI applications. Users are already questioning the underlying mechanisms of AI systems in everyday use. They criticize the results of digital assistants, for example, and demand explanations. They are passing this attitude on to their children, the so-called AI Alphas: the generation growing up with AI as a matter of course. The children and newborns of today will, in future, expect a bias-free technology they can trust.

Research also plays an important role in spreading trustworthy AI systems. Critics have long warned of the negative impacts AI applications can have. At universities, in think tanks and among government advisors, researchers are working to address the challenges of the black box.

When users can trust algorithms because they deliver correct results, the customer base grows.

Companies share responsibility for AI-driven decisions

While developers of AI applications face the task of designing them in a more trustworthy way, companies that deploy AI-powered programs also share responsibility for ensuring trustworthy use. If AI applications refuse credit to founders, for example, without considering the wider context of their business plan, not only does the financial provider’s reputation suffer, so does its profitability. Only when users can trust algorithms because they deliver correct results does the customer base grow.

Wrong decisions by AI can have even more serious consequences in medicine and the wider healthcare sector. Patients, doctors and nursing staff form a highly sensitive environment. Inadequate data leads to misdiagnoses and can even have fatal consequences. A trustworthy AI that supports doctors in their decisions and takes all circumstances into account can, by contrast, deliver truly personalized treatment for patients.

AI Trustability and more trends at a glance in our Megatrend Map

AI Trustability is one of a total of 120 Macro Trends from our current Trend Universe, which is grouped into 17 Megatrends. With our free Megatrend Map, you get an overview of our entire Trend Universe, including explanations and our analysts’ assessments of mainstream adoption.