Listened to the interesting Geopodd on KKrVA's blog about AI and geopolitics and have some additional comments. There are mainly two additions I would like to make. First, the need to normalize AI, that is, to use AI as any tool – not without a critical eye, but constantly, much like a calculator. Second, to realize that AI is already used intensively in both geopolitical analysis and in communication and influence campaigns.
That AI has been “weaponized” on the battlefield is already well established. Millions of sensor data are collected in real time, verified, sorted, prioritized, coordinated and finally fought in good order, even after human approval. AI played a central role in the US campaign in Iran and, according to the Pentagon, contributed to very good results. This provided a short-term and limited target picture, if you want to be critical. I wrote about this earlier this spring on KKrVA's blog.
The fact that the information area has also been “weaponized” is perhaps less well known and talked about, but more widely used and probably with greater effect than the battlefield variants. Automated AI agents analyze, plan, specify targets and carry out propaganda and deception operations on a large scale, on behalf of both state and criminal actors. Deepfakes that are hardly detectable, phishing messages and scams, fake news articles, fake statistics, fake analyses, etc. In 2025, the time from idea to operation on the Internet, for example, has in many cases shrunk from hours to seconds.
In a world where both reality and truth have become relative, our ability to question and expose these images of reality and “facts” is being put to a severe test. Even our authorities, such as the Psychological Defense Agency, are having trouble monitoring, catching up with, and counteracting this stream of influence. Even our central media outlets are having trouble sorting through this stream of disinformation. The crisis in the Middle East, for example, provides good evidence of this.
AI is therefore not a question of the future. AI is an advanced but established everyday tool for both public actors and organizations, and we must realize this as individuals both privately and in our professional roles. Not utilizing these tools means a major handicap; they have inherent strengths that are impossible to compensate for through intelligence and general wisdom. Then, of course, wisdom must be used to sort, verify and evaluate AI's data before they are used. The power of the information that cross-referencing different LLMs such as Claude, ChatGPT, Grok, Perplexity, etc. provides cannot be overestimated and takes OSINT, for example, into a whole new dimension.
When I obtained some background, by cross-referencing AI, for my previous text on KKrVA's blog about the integration of Palantir MSS in Sweden and our need for a national secret cloud service, I received information that I chose not to use. I thought it was too sensitive. For example, a fairly detailed timetable etc. for implementation and integration. Whether it was correct or not, I of course do not know.
Use AI as an information service and basis, but check sources and test run against other AI – use the tools. It is impossible to read through the thousands of pages of basis that AI rakes into its “information bag”. Then I recommend trying using “vibe programming” to create your own agents through, for example, the Swedish service Lovable. Months of Python programming can be reduced to a few minutes. If nothing else, you learn a lot about how AI works and can be used through a little “home carpentry” in the area. I am an economist, not an IT engineer, and my knowledge of programming is extremely limited.
The gist of what I want to convey about AI is not about the future, it's about now, so start using the tools. Development is dizzyingly fast; in a month, progress is made that previously took years. A year ago, people were talking about AGI (Artificial General Intelligence) becoming a reality in fifteen to twenty years. Now some are talking about two to three years.
But remember – until AGI eventually arrives, we are the ones setting goals and planning for the future.