AI Is Stealing Your Intimacy!

I watched the discussion on Finnish Broadcasting Company´s (Yle) morning show “Ylen Aamu” on 11.2.2026 before 7 a.m. The topic was “The Effects of AI on Data Security,” and whole session seemed poorly prepared and was awkward to watch. Did you happen to share this experience with me? (not sure if link provides you EN subtexts). In my view, the journalist’s perspectives were superficial and exaggerated, and the experts present were pushed toward portraying AI as guilty of data collection. The experts included representatives from the Cyber Security Center and the University of Helsinki.

Both experts made it clear right at the start that AI doesn’t really collect information about us; most of the data has already been collected. The journalist could no longer blow up the script and ended up pushing the discussion to the end with the same fear-inducing, pre-planned questions. The angle chosen was data security, but the exaggeration went overboard, and the discussion mainly created fear for those not familiar with AI. Some of the threats mentioned seemed to be years old. People were encouraged to avoid AI, even though it inevitably becomes a citizen skill, just as “computer literacy” once was.

The interview reinforced my perception that we’ve been giving away even intimate information at an accelerating pace for the past 15–20 years—at the checkout, on social media, or through any online service. As location data, search history, purchasing behavior, social media content, device information, and browsing behavior.

For years, our behavioral data has been sold by IT firms to advertisers, and using that information, we’ve been manipulated every day with targeted ads drilling into our preferences or addictions so that our money ends up in the desired accounts. We constantly see less content related to our friends in our social media feeds as algorithms become increasingly commercial. See also Netflix: The Social Dilemma. This thing shouldn’t be blamed on AI.

AI models combine data, recognize behavior patterns and writing styles, make predictions, and can apply information more efficiently through computations. So what truly new, original data do they extract from us? Perhaps more voice samples, biometric data like facial or fingerprint information, chat content (prompts), or sensor data from smartwatches or health apps—if that hasn’t already leaked.

Of course, language models continuously produce increasingly realistic-sounding text because they’ve been trained to understand and generate language humans have already used. And of course, the latest AI models, like ByteDance’s much-discussed Seedance 2.0, generate astonishingly realistic-looking videos and movies that are now almost impossible to distinguish from traditional authentic scenes, which already include trick photography, effects, and digital editing.

Responsibility will initially rest at least partly with the consumer, who chooses whether to believe in the authenticity of an image, text, or video. Some will believe, some won’t, and some won’t care or be interested if the material is pleasing, enjoyable, or useful. Perhaps one day regulations will emerge requiring authorities to mandate certain AI identifying information. Maybe a classification system will be developed indicating the degree and type of AI involvement. Now as this came to my mind, I naturally asked AI whether anything like this is currently being developed:

ChatGPT: “The European Union AI Act is a major regulatory framework concerning the use and transparency of artificial intelligence. It does not directly provide an ‘AI degree’ or numerical rating for, say, movies, but it sets requirements on how users must be informed about the use of AI systems and what kinds of risks are managed in different categories. This regulation is one of the first attempts to create a legal framework for responsible AI use, but it does not establish a media or content classification for media products.”

It’s certainly important that AI-generated content can be distinguished, or if it can’t, that this is communicated to the consumer. A critical need, of course, is to keep traditional research-based facts separate from AI-generated interpretations or entertainment. Genuine art and culture, where the user forms a relationship with human creators, must remain distinct from fictional AI fluff.

Then a harder question: to what extent can verifying an individual’s skills and professional competence in the workplace rely on AI-generated content in the future? Probably quite a lot! Educational background—such as degrees and certifications—and work history obviously need to be authentic, but what is then built on top of that?

Processes are becoming increasingly complex, and cause-and-effect chains are harder to see with the human eye or to grasp. The demand for efficiency has already reached a level where no one has time to spend hours or days investigating a single problem. There’s too much data, and documentation is laborious. In everyday problem-solving, many decisions are already based on theories and probabilities, constructing a plausible logical narrative, and teamwork.

So, in the absence of exact facts and evidence, the explanation that offers the most plausible story wins. If you can also structure that story logically within the given time frame, it becomes “truth.” Few of us are more talented than AI at this, and you yourself can define in your prompt what should be treated as boundary conditions—so-called facts of the story.

Do you really believe AI is stealing your intimacy? I think AI shouldn´t provoke us, and it can’t be feared. Anyone benefits from understanding it, regardless of the threats painted around it. While limited perspectives like Yle Aamu’s push us in one direction and our data is already widely processed by AI models, we can simply take advantage of it to make our own lives easier.