Two-thirds of Dutch workers do not remove sensitive data before pasting it into an AI tool
In the Netherlands, 5.2 million people now use AI at work, only 29% have had any training, and just 34% strip sensitive information out before pasting it into a tool. The share who say their employer has clear AI rules fell from 57% to 42% inside a single year. Fieldwork July 2026, published 25 Aug…
In the small nation of the Netherlands, 5.2 million people already use AI at work. Among them, only 29% have had any training, and just 34% strip sensitive information out before pasting it into a tool (think ChatGPT, Claude, Gemini). The true problem here is: DO THEY USE THE COMPANY ENTERPRISE ACCOUNT, OR A FREE ACCOUNT? Dutch AI use went up while the share of people who know the rules went down On 25 August, Frankwatching published the fifth wave of the Newcom AI-Monitor, written up by Fleur Zick. It surveyed 3,122 Dutch adults aged 18 to 65 in July. Neil van der Veer of Newcom summed it up plainly, that employees have made AI part of their daily work themselves. A few numbers at a glance: 5.2 million Dutch people use AI at work, up from 2.8 million in 2024 57% say it is now part of their routine 29% have had any training or instruction, and 47% say none is available 64% feel personally responsible for what the tool produces 59% check the facts it gives them 34% strip sensitive information out before pasting it in Are we then expecting the output to be showing signs of mis-use? Perhaps some data mining is in the mist? The governance is bad Nobody scrapped their AI policy this year, not exactly. There will be some not so jolly folk who would say AI policies have been removed. The likeliest explanation is a moving denominator. As AI use spreads out of the governed centre of a company and into the corners, the share of users covered by a policy falls even while the number of…