Impressions from Testnet’s Najaarsevent 2025

Twice a year, a test event is organized by Testnet in the Netherlands. The autumn event of 2025 was another great edition with no less than 3 keynotes, several workshops, and many good talks about software testing.

The afternoon session started with a keynote by Darren Richardson. His keynote was about the Dead Internet Theory. He himself is a security specialist, so he views everything from a security standpoint. With AI, we currently face the problem that there are many things on the internet today that are fake and generated by AI. More and more phishing attempts are also being created with AI.

Another issue is that there are many bots active that post comments under posts. He observes that bots are starting to respond to posts made by other bots. So, this theory suggests that in some time, only bots will operate on the internet. This is also what Zuckerberg, the founder of Facebook, has already said, that more and more bots are coming that will create content. His company must keep the users’ attention because Facebook is a large attention machine. Darren ultimately concludes that AI is not ethical. According to him, Corporate Social Responsibility has failed. Companies continue to use data that is not meant for their profit model.

Fortunately, he also offers some solutions after explaining the problem. We should use AI, for example, to assist us, and not to replace us. The main point is that everything created with AI must be labeled. You should indicate, for example, which AI agent it was created with, which model, when, and so on. It is a whole data structure with all relevant information. In Europe, we are making good progress, but the legislation is still slow. But he says, you shouldn’t complain about Europe being slow, or because we can’t just use models and data. They are doing something to protect citizens against this. Because if you complain that you can’t use it, you shouldn’t later complain that they are really taking our jobs.

After that, I went to a presentation by Marianne Duijst, a Quality Coach. The title of her presentation was “Quality Culture Foundations: Are you traveling on the software highway yet?” It was a very inspiring talk about her experiences with several teams. She pointed out that in a company, they had 2 teams with old software and new software. But after a while, the 2 teams no longer knew what the software did. The new team had to replace the old product.

The business people were also not really satisfied because the new software was all in English, while the old one had everything in the database in Dutch. She then introduced a demo session every few weeks where the teams showed each other what they were working on. She actually sparked communication to try to create a quality culture among the teams. Through this communication, a dynamic emerged where everyone started thinking. Testing began alongside requirements. The business was suddenly also involved in the process. She started small, but it ended with a significant change where there was continuous feedback, learning moments, and above all, much happier team members.

The presentation by Rob van Steenbergen and Bram van den Berg was also worthwhile. It was not about AI but about the Agile Test Model. The Agile Test Model (ATM) provides a visual overview of all testing activities and shows where they take place in relation to the customer, concept, and product. It is not a process model but a static model that helps testers, developers, and the business understand, discuss, and position testing.

They discussed the relationships and interactions needed to see where you can test or where you can fulfill quality. It turned out to be everywhere. We see that testing impacts everything. We talk to the customer, we check whether the build process has produced the right product, and then we check whether we have made what we agreed upon with the customer. By using this model, you can check in your team where there are still gaps in the testing and quality processes.

Then I entered the presentation of Tom Heintzberger. They developed a product that can analyze test output using AI. The tool can help us identify the cause when a test fails. Is it the test case itself, or is it a bug in the product? Have there been similar problems before? What is the root cause of the problem? The tool can also provide historical insight into the test. Is that test stable? Does it fail often? But also whether the tests are relevant. Perhaps there are still tests that already do the same thing.

There are still some risks associated with this, such as hallucinations. This mainly occurs if your model is not kept up to date. So that is apparently important. Also, regarding data security, there are still some risks. Are we allowed to put that data into an AI? Like with any model, there is also BIAS or drift. We must therefore always check the output as humans. Never blindly trust that everything is correct.

At the end of the day, Dick Tracy was allowed to give his keynote “We Are The Testers: You Will Be Assimilated.” It was about the fact that we have an unbalanced team. Usually 4 developers and only 1 tester. What if we reversed the roles? 4 testers and 2 developers, for example. Then we could examine much more as testers because the tasks of a tester are so broad and so much that you can never do it alone. He had developed a model for this and analyzed it completely. At the end of his talk het told us to try it.

I am not sure if we can though, we already have not that much budget on staffing.