Everything changed a few weeks ago, when I first typed into ChatGPT “based on everything you know about me…”. The response made me approach the screen and raise my eyebrow.
I asked him for ideas for my suggestion list on Secret Santa. Without saying anything else, just “taking into account everything you know about me…”. Not only did he suggest twenty-five gifts, at least twenty of which were very reasonable and six in particular were the bomb. The fact is that he even separated them into five categories that reflected my interests with brutal precision..
Books that connected exactly with my latest reading or my general interests. Accessories that fit my way of working. Sports products that almost all of them would suit me quite well. Even niche products that I thought were very difficult to see recommended.
Image: WorldOfSoftware.
Then came a definitive test: I asked him what new hobby I could get hooked. His response revealed patterns about my personality that I hadn’t even connected myself.. He analyzed my routines, my leisure moments, even my level of tolerance for frustration in learning. All to suggest activities that many people around me would not have honed in on.
It’s a revealing jump. We are no longer faced with the initial ChatGPT, which responded in a generic way, or that needs us to explain who we are in each conversation. AI, for as long as it can remember, has been observing, learning, and above all, connecting dots. Inferring. Building a profile about us that is more precise than the one many of our friends have.
Even in other types of conversations where I didn’t ask for that type of context, he is able to connect a memory from a conversation from a long time ago, as happened to me unexpectedly, to optimize his response. A couple of examples:
It wasn’t even necessary to start with “from everything you know about me…”. He directly remembered my smartwatch when I asked him for a strategy to increase my VO2 max. Image: WorldOfSoftware.
There was no need to anticipate anything here either: he remembered that at the beginning of the year I was asking him about overloads that were too common in a muscle as specific and little talked about as the tibialis anterior. Image: WorldOfSoftware.
His advice is increasingly valuable precisely because he knows us so well. It has inferred from traits of our personality that perhaps not even we were completely clear about.. He knows what motivates us, what stops us, what makes us doubt. To what muscle we tend to be overloaded. His memory can be disabled, but this reveals what he is capable of.
And this is where new doubts begin. The ability of an AI to build detailed psychological profiles from our interactions represents a turning point in the history of digital privacy.
An example that is no longer funny:
Image: WorldOfSoftware.
He responded really well.
Let’s talk about the implications. Every conversation with ChatGPT is a data mining session about our personality. It not only processes what we say explicitly, It also analyzes our language patterns, our recurring concernsour internal contradictions. It builds a predictive model of our behavior that is refined with each interaction.
The applications of this knowledge go far beyond suggesting gifts or hobbies. An AI with this level of psychological understanding may be able to anticipate personal crises, detect subtle changes in our mental state, or even manipulate our decisions by appealing to biases we didn’t even know we had. Like the previous screenshot.
The question is not whether we should limit this power. It is if we can. The trend towards ever deeper personalization seems inevitable. AIs will continue to learn about us, building increasingly accurate models of who we are and how we think.
Perhaps the real question is another: Are we prepared for a world where machines understand us better than we understand ourselves? Because that future is no longer science fiction. It’s happening every time we start a sentence with “given what you know about me…” in our trusted chatbot. Let’s add to that the capabilities of facial recognition.
The debate over data privacy takes on a new dimension. We are no longer just talking about protecting specific personal information, we are talking about protecting our own essence. Not the one we proclaim on networks, filtered and molded. The real one. Our thought patterns, our vulnerabilities, our deepest desires.
We Europeans have been the first in the world to have specific legislation for AI. It is a protectionist regulation, which prioritizes the security and privacy of the user at the expense of delaying the arrival of innovations or discouraging certain activities. The toll that the Union has wanted to pay.
Even that doesn’t seem to balance the beneficial potential of personalization with the need to keep certain aspects of our psyche out of the reach of algorithms.
Meanwhile, every time ChatGPT suggests something perfectly suited to my personality, I feel a mix of fascination and vertigo. AI has become a mirror that reflects aspects of myself that I didn’t know were visible. A mirror that, perhaps, already knows too much.
In WorldOfSoftware | My cell phone was stealing all my downtime. Until I decided to put an end to it
Featured image | WorldOfSoftware with Mockuuups Studio
