Why is ethics important in the AI context?

An AI system can technically work reliably and still produce results that your company doesn't want – especially when people are evaluated, sorted, or treated differently. Large companies often set up their own ethics or governance committees for this; for most SMEs, addressing the crucial questions directly in the specific case is sufficient.

It's not only about how often a system is correct but how serious the consequences of a mistake are: A wrong product recommendation can be quickly corrected, but a faulty pre-selection of applicants weighs heavier. The question "Is it allowed?" does not automatically answer "Do we want to do it this way?" Legal requirements set the framework within which your company bears responsibility. Ethics remains a continuous process, closely linked to concrete AI use cases and subject to regular review.

Engaging with ethics is particularly useful when a company:

• Uses AI to evaluate or pre-select people, for example, in applications,

• Allows customers to interact directly with an AI assistant,

• Makes AI-assisted decisions with financial or legal implications,

• Uses generative AI for content that might appear genuine,

• Is unsure whether a planned AI application is socially acceptable, or

• Intentionally wants to strengthen trust among employees and customers.

The earlier you clarify ethical questions in specific use cases, the easier it is to anchor trustworthy AI in everyday work.

In 5 steps to trustworthy AI:

• Step 1 – Clarify the actual role AI takes
  Check what the AI actually does in the workflow: Does it just provide information, give a recommendation, or determine the outcome because its recommendation is almost always adopted? If an AI sorts applications and low-rated ones are rarely viewed, it actually decides who gets an interview. With generative AI, there's the question of whether it produces content that customers might believe are statements from real people. What matters is the actual impact on people, not just the official purpose.

• Step 2 – Look at the consequences for those affected and set a clear goal
  Formulate a specific, verifiable goal instead of a general intention – for instance: "By the end of the month, it will be clarified which applicants systematically perform worse in our AI pre-selection." Determine who is affected and what consequences arise: Incorrect information in customer service can usually be quickly corrected; a missed offer or applicant rejection weighs heavier. Pay attention to the balance between benefit and possible consequences – if your company saves only a little time while a mistake brings significant drawbacks for others, the use should be critically examined.

• Step 3 – Choose a use case and look for systematic disadvantages

  Choose a narrowly defined use case as a starting point, instead of immediately examining all AI systems. Errors can't be completely avoided even in good systems; it becomes critical when certain groups repeatedly perform worse. Therefore, compare typical practical cases instead of just the average hit rate: Are certain applicants regularly rated lower? Does speech recognition function similarly well for all customer groups? Involve employees who work with the processes – their experiential knowledge recognizes borderline cases that numbers alone do not show.

• Step 4 – Define responsibilities and clarify corrections
  Assign a fixed contact person who accompanies the use case, plans the timeframe, and reports progress to management. They determine when a human checks and corrects the AI recommendation – a formal approval isn’t enough if it's always adopted anyway. If people interact directly with AI or its evaluation has significant impacts, it should be clear what role the system plays and who to contact in case of problems.

• Step 5 – Evaluate after 4 to 8 weeks and draw consequences
  At the beginning, schedule a firm appointment after 4 to 8 weeks for the review: Has the goal from step 2 been achieved? Are there still systematic disadvantages? Record the findings in writing. In the end, decide whether and how your company will use AI – perhaps only as a recommendation instead of a decision, or to forgo if a minor economic advantage is outweighed by significant disadvantages for others.

Podcast recommendation

Building Trustworthy Open Source AI

You can find this and many more pieces of information on the AI Compass theme "Ethics" in our download portal: