The next frontier: Artificial Intelligence we can really trust.
Artificial intelligence (AI) is currently very successful. Three reasons are responsible for this. First, the enormous progress in the field of statistical machine learning. Second, the availability of large amounts of training data. Third, the increasing computing power available. For certain tasks, algorithms can achieve very good performance. These even exceed human levels.
Explainability and Robustness are the ingrediences of trustworthy AI
Unfortunately, the most powerful methods have two weaknesses. On the one hand, their complexity makes it very difficult to explain why a certain result was obtained. On the other hand, they lack robustness. The most powerful machine learning models are very sensitive to even small changes. Perturbations in the input data can have a dramatic impact on the output. This can lead to completely different results. Poor data quality is one of the factors responsible. We do not have the i.i.d. data we expect.
AI in domains that impact human life
The use of AI in domains that impact human life (agriculture, climate, health, …) therefore leads to an increased need for trustworthy AI. In sensitive domains such as medicine, where traceability, transparency and interpretability are required, explicability is now even mandatory due to regulatory requirements. One possible step to make AI more robust is to combine statistical learning with knowledge representations. For certain tasks, it may be beneficial to include a human in the loop.
The human-in-the-loop
A human expert can – sometimes, of course, not always – bring experience, domain knowledge, and conceptual understanding to the AI pipeline. Such approaches are not only a solution from a legal perspective, but in many application areas, the “why” is often more important than a pure classification result. Consequently, both explainability and robustness can promote reliability and trust and ensure that humans remain in control, thus complementing human intelligence with artificial intelligence.
Read the paper:
Andreas Holzinger (2021). The Next Frontier: AI We Can Really Trust. In: Kamp, Michael (ed.) Proceedings of the ECML PKDD 2021, CCIS 1524. Cham: Springer Nature, pp. 1–14, doi:10.1007/978-3-030-93736-2_33.
