The topic of artificial intelligence is becoming increasingly relevant. For this reason, the Fraunhofer Institute is also increasingly concerned with the possibilities of AI. In its latest blog post, the Farunhofer IAO has now published a rapid test, which should help entrepreneurs to assess which activities would be suitable for AI and which should remain purely manual.
Implementation test part 1: Artificial intelligence likes routine
As Fraunhofer writes so beautifully and as the practical examples show, artificial intelligence loves nothing more than routine. Through machine learning, thinking machines gradually learn about certain recurring behavioral patterns and are able to interpret them. This is precisely why tasks that involve a high degree of routine are usually particularly suitable for automation. The quick test 1, which only allows “Yes”, “No” or “Don’t know” as answers, helps with an initial process overview.
.
- Does the work task occur in the same way, e.g. by frequently working on the same tasks?
- Are the decisions you make to perform the task defined by rules?
- If rules are defined: Are these rules accessible and writable?
If all questions are answered with “Yes”, the use of an AI is optimal. Even with small deviations from the routine, the artificial intelligence can react and adapt to these tasks. But, as always, the AI cannot take over all human activities here. As mentioned in the current podcast, it is only a tool to relieve the employees and to enable them to concentrate on the essential.
Implementation test part 2: High requirements make immune against automation
In contrast to routine work, demanding tasks are rather difficult for automation using artificial intelligence. Even though AI learns adaptively and can react to innovations, it reaches its limits when tasks are too complex.
As in the first part, there are only three possible answers for the corresponding process you want to consider…
- …high thinking requirements, e.g. because you are dealing with complex and knowledge-intensive topics?
- …high interaction requirements, e.g. because you have frequent and long conversations with colleagues or customers?
- …high creative requirements, e.g. because you often develop solutions for new kinds of problems?
- …high demands of the holistic situation assessment, e.g. because you use a variety of information and sensory impressions for your decisions ?
- …high demands on the use of intuition/ instinct, e.g. because you make quick decisions in dangerous situations ?
- … high demands on cultural and ethical standards of evaluation, e.g. because you work a lot with foreign people?
- … a high degree of freedom of action, e.g. because you have discretionary powers when making decisions?
A single “yes” is already decisive here. If the task becomes too complex, it is currently not suitable for artificial intelligence and should not be automated.
These and other Fraunhofer activities, such as their AI toolbox, are an important step towards bringing understanding and transparency to artificial intelligence technology. Only when all managers, employees and partners can circumnavigate the pros and cons of artificial intelligence will they be able to get the most out of their work tasks and take full advantage of the possibilities of artificial intelligence and automation.
