AI can help us understand reality, but it cannot replace direct experience with it

In Santander X Explorer, the use of AI is summed up in the phrase above. When used well, artificial intelligence is a very powerful tool for progress. But if it is used solely to “get the job done,” a very valuable learning opportunity will be lost.
Without this learning, every step could lead to a wrong path, since they will not have occurred in the real world, and most importantly: will lack human judgment. Artificial intelligence will be making decisions for us, and its resolutions (we hate to say it) will be useless, since they will not be built on real evidence.

Can evidence be fabricated and work from there?

Going out into the world can be a bit intimidating. The temptation to invent a customer or a purchase intention is there, but everything hypothetical or untested against reality will inevitably lead to a very nice and convincing response… that will still be a supposition, not evidence. AI must be used to prepare for this step into the world and to analyze it afterward, rather than to avoid it.

The more real evidence are provided (results, responses, behaviors, data), the more useful it becomes in finding patterns, contradictions, and insights. When the signals collected (generated by people and real behaviors) are used for AI to form hypotheses, design experiments, or analyze results, there is learning. It offers a direction that represents progress.

The information we gain from interacting with reality cannot be replaced by AI. 

Facts vs. ideas

The way we ask artificial intelligences conditions their response, which is why it is useful to ask them to identify what we take for granted, unsupported claims, logical gaps, or weaknesses in an idea.

The alternative explanations AI may propose will open our eyes and lead us to new questions and insights, in addition to strengthening the project. We will avoid falling into self-deception.

In Explorer we propose “training” AI to find objections to reasoning or question conclusions. We will learn to distinguish what we actually observe, what interpretation we make of it, and what we imagine. AI can actively help to make this separation.

The decisions remain ours

AI helps us think, but it must not decide for us if a hypothesis is validated, if iteration is necessary, or if it is best to abandon that specific project and start over. That resolution must rely on the evidence and, above all, on our own judgment.

Therefore, after using an artificial intelligence, we must ask ourselves: “Is there learning here, or is this just another convincing story?. If we have gained new knowledge about the project, AI has amplified the process. If it has only made what we already thought more believable, will be fueling self-deception.

The function of AI is to reveal a learning insighta

AI must help better understand an idea, ask better questions, and make more conscious decisions. If it does the work for us and we learn nothing, it is not fulfilling its purpose.

Do you want to learn how to start a business in the AI era? Join Santander X Explorer: https://explorerbyx.org/lk/bNp9