


European Union Law Firm

European Union Law Firm
June 20, 2024

——————-
Types of artificial intelligence (AI) models include large language models (LLMs) such as ChatGPT, Gemini, Copilot, and Claude, to name the most popular.
These models operate in the following way: The AI absorbs a large amount of images, texts, and other human creativity and, through mathematical, statistical, and probabilistic algorithms, uses that creativity to generate other text, images, and, more broadly, other content.
These types of AI seem to be realizing real industrial assembly lines of thought.
In essence, we have transitioned from the phase of industrializing products and services to the phase of industrializing thought and, in particular, intellectual work.
From this introduction, at least three immediate considerations seem to follow.
One very basic consideration is that AI, acting as an assembly line for human creativity, presents several challenges regarding data origin, including issues related to intellectual property and the processing of personal data.
The second is that in order for AI to work properly, the needs, intentions, and individuality of the AI user are essential.
The third is that a dependence on AI or acquiescence of its responses could, over time, decrease our ability to innovate and produce critical thinking.
At a superficial glance, AI doesn’t operate differently from humans.
It learns to generate, just as we do.
Upon closer examination, a fundamental difference emerges.
Our need to learn and generate is filtered through our individuality, which is a combination of emotions, experiences, and culture that has developed and evolved over a particular historical period, thus belonging exclusively to our personal history.
Unlike AI, we do not generate to provide answers but to create solutions to our needs.
So AI responses can only produce solutions if humans are able to make sense of them with their own needs, intentions, and individuality.
These aspects, independent of training problems, are also the source of some of the limitations of AI, such as hallucinations, the production of false information, difficulties with structured data, toxic or harmful discourse, etc.
Indeed, to ensure that AI generates a solution, it is necessary to give a prompt that controls the tone of the response and specifies our desires as accurately as possible, thereby imparting intentionality to AI’s output.
Furthermore, LLMs, being designed to generate from human creativity, cannot produce outputs different from the human creativity they take as input.
This is why they cannot perform calculations on their own, and this function is left to technical solutions.
From all this, we come to the third consideration.
If AI responses do not go beyond the human creativity already produced and are the result of a “mechanized” digital process, the risk over time may be that we will mostly be content with preconstituted responses that do not lead to giving or creating new solutions or innovating.
The other risk is consequential: from aquiescence to AI responses comes single-mindedness and the weakening of our critical thinking.
In fact, AI responses have only one source: the AI itself. And it is a source that is objectified because it assimilates and generates without individuality.
In this regard, the comparison with the assembly line and the difference between industry and handicraft can help me explain better.
In our homes we have fewer and fewer handcrafted (and therefore unique) pieces and more and more pieces of furniture that we can find in the homes of many, if not all. They are undoubtedly useful and practical products, but our homes are becoming less and less original, more and more flat and standardized, with the consequence that our sense of beauty is also less and less pronounced.
Only here it is not about furniture but about knowledge and thinking.
From another perspective if we start from the idea of the industrialization of thought, we must also conclude that work will probably not be the same and that we should probably already invent another kind of economy and society. We could also consider that if we leave intellectual work to AIs, sooner or later we could not have the same inventive capacity because inventiveness depends on work and the need to overcome obstacles.
I would like to emphasize that AI in my opinion should be seen as an opportunity to expand our horizons and do things that we would not otherwise be able to do, but I also believe that it should be used, regardless of any ethical or legal solutions, with awareness and giving it its proper weight.
I leave it up to you for any further consideration and the opportunity to make your own contribution to the discussion on blog or linkedin.
Leave a Reply
You must be logged in to post a comment.