OSINT Is Still a Thinking Game - Surviving AI Without Losing Tradecraft

Niko Dekens, a long-time Dutch government OSINT practitioner, argues that AI will not replace analysts but can wear down their critical thinking. He describes how chatbot answers look convincing yet are often wrong, and how synthetic accounts, fabricated documents and AI-edited images are increasingly seeded so that search engines and language models surface them. The question, he suggests, is no longer only whether a source exists but whether it was made to be found.

He illustrates the risk with examples. Students geolocated an AI-generated Amsterdam street and accepted it without checking a map. A Russian-language Telegram phrase was translated as an imminent attack when it meant the opposite. A sudden drop in shipping volume was read as financial distress until a port strike turned out to explain it. He also covers exposed chat histories that reveal investigative focus, and online grooming that pushes young people toward violence through closed messaging channels.

His defense framework has four habits: read the raw source rather than the summary, force a second hypothesis, separate speed from confidence, and audit dependence by labelling AI-assisted content in reports. The practical takeaway is to keep reasoning visible, verify top results by hand, and ask what evidence would change the conclusion.