The intelligent system that always lies


A company is developing an AI-based system for verifying news and facts on the internet.

The system is supposed to automatically detect whether a report is true or false and protect users from misinformation.

After launch, however, users report that the system often gives incorrect assessments and sometimes even classifies obviously true facts as false.

The development team wonders: How can it be that an intelligent verification system, despite large amounts of data and modern algorithms, constantly makes mistakes and seemingly "lies"?


Question:
Why can an AI-based verification system, which relies on extensive data and complex models, repeatedly make false assessments and even declare true information as false?

Solution follows tomorrow.