AI – business model and education

The world of advertising today works in “push” mode, that is, they try to study the user as much as possible and push him/her solutions even if he/she hasn’t thought of them yet. This model has two main problems, the first – it is a statistical result and therefore not necessarily suitable for the specific user, and the second – to learn about the user the system needs to intrude his privacy.

The artificial intelligence solutions make it possible to reverse the equation and switch to “pull” mode. That is, the user will give all his details (including the most intimate) to a “digital personal assistant” under two basic assumptions – this information will not be shared with any person or entity, and the security system will ensure that no one can break into this database (difficult but doable basic assumptions). With this information and the existing artificial intelligence capabilities (and those that will be developed soon) the user’s “personal assistant” knows how to match the user with exactly what is required. Let’s say I’m interested in buying an office chair for my study. The personal assistant knows the budget, the color preferences, the style and the fit with the existing furniture, my ergonomics and the previous furniture choices I made and will recommend a number of chairs that suit me with great precision. How will Google profit (assuming they developed the personal assistant)? They will charge me a monthly subscription fee and charge a fixed percentage from the supplier for the purchase. The personal assistant will give feedback to the network regarding the service, the quality of the product, the maintenance throughout the product’s life cycle (not only during the purchase). The provider, for its part, will create as detailed a content as possible on the subject so that the personal assistant can make the recommendation decision. In this way we provided personally accurate answers to the user and solved the privacy problem.

How is all this related to education (after all I usually write about education)? Well, anyone who has played with GPT and the like cannot escape the feeling that these are general answers, like a guide to the good life that knows the answers to everything. It’s great when I need to do homework, check how to fix my car or develop an app. In all of them a general answer is sufficient, except that a general answer is less good in education when the adjustment to the specific student is mandatory. In order to teach, it is necessary to know the student’s abilities and preferences, style and motivation, as well as the student’s prior knowledge. For this purpose, the personal assistant must be given very intimate details about the student and I don’t know many parents who, in the current model, would agree to this, so we are left with a tool that serves the student (to prepare homework) but does not serve the teacher and the education system. Therefore, a “pull” model that comes together with a privacy and safety protection contract can be suitable for the education system (and for Google as well).

If I were Google, I would start working today on a conclusive contract that mainly protects the users and allows to study us in depth, and to look into a quantum safety system or something similar.