Everywhere you look, people are saying that AI, increasingly used in the workplace, makes us more productive. We outsource low-value tasks to them, generate a first draft, an imperfect raw clay, and then we refine this crude material into a perfect deliverable. In short: less time spent, yet the same or even more added value. AI handles the initial, tedious work of data collection, analysis, correlation, and generating a first response. Humans add the finishing touch, the fine-tuning, the final brushstroke, allowing us to produce in record time what used to take ages in the pre-AI world! That's one way to see it! Thanks to this well-known heuristic: the more sophisticated the technologies become, the more they take over menial tasks that were assigned to us before (thus freeing up our time to create that famous "added value").

But if we follow this reasoning, we're mistaken. The rise of so-called "generative" AIs changes the game. The belief in human liberation and enhancement through technology no longer applies. Why? What's changing?

Generative AI is not just another productivity technology! It "generates"! This little word, seemingly harmless, is actually loaded with meaning (and consequence)! It means that it learns (deep learning) from the data it is trained on. Based on this, it can create its own interpretation and response (thus fabricated) when queried. Mid Journey, Stable Diffusion, Dall-E, these visual generative AIs do not merely assemble pieces of images found in their database. They learn to do "like." And the images they produce are created ex-nihilo! Hence the magic. There is virtually no response a generative AI cannot construct.

So yes, the generated responses are still imperfect, and humans need to tweak them to reflect our image. Chat-GPT sometimes speaks in too "doctoral" a manner; the text needs a bit of "roughening."
But soon, the outputs will surpass those produced by humans (in relevance, realism, mimicry).

Doppl, a French startup, is developing a "digital twin" AI, meaning a generative bot trained on your personal digital data, capable of "resembling you" and thus generating content "like you." Imagine the convergence of this technology (individual micro AIs) with global generative AIs (developed by Big Tech): you'd have an "alter ego" working for you.
Black Mirror?

To wrap up our initial topic, we need to let go of the idea that AI is the “factory” and humans are the "brains." There is, and always will be, value generated on both sides. A draw, so to speak? In any case, there will be a match, and the competition looks fierce.

Moreover, what is Neuralink if not Elon Musk's ambition to enhance the human brain directly with an AI layer via bio-implants. Why do this? Simply so that in the upcoming match, tomorrow's human can compete on almost "equal" footing and continue to stand a "small" chance against machines...