AI, a free all-you-can-eat buffet

They suck every expression they find in their path. They are silent and invisible artifacts that ingest as much human language as they can read, hear, or see… perhaps with these words that sound like a riddle, someone from the past could describe what Artificial intelligence represents and does. I’ll tell you more about the monster right away.

Indeed, ChatBots, those Artificial Intelligence machines, are tireless and voracious. Day and night, they keep copying and processing any datasets they find online. Messages and comments on social networks, website content, open books, university theses, medical reports, newspapers, open forums, Blogs with B and V like this one…

But it’s not just text; it’s also images, videos, sounds, and music. Like sponges, they absorb anything human language expresses, including jokes. Anything goes if it is language, which, in addition to being verbal, is visual and audible.

Well, this flow of information we leave daily on the network becomes the food with which we feed those peculiar “gremlins” called ChatBots. As a result, these AI conversational robots are getting smarter and faster in intuiting, predicting, and simulating what a human will answer to this or that question. Yes, indeed, we are feeding the monster, and we do not exactly grant informed consent nor get paid to supply that valuable raw material that is human knowledge.

In this way, OpenAI’s ChatGPT, Microsoft’s Bing Chat, Google’s Bard, and other Generative Language Models are “hacking” the operating system of the most powerful computer in the universe: the human brain. This statement, which belongs to the writer Yuval Noah Harari, does not seem exaggerated to me because the stealthy procedure of these robots consists of absorbing incredible amounts of human language to learn to simulate it.

From a single expression, they can predict which words should follow. There is no reasoning, but the predictive and imitative ability is so high that it seems human, which is precisely the danger. There will come a time, and perhaps we are already there when we do not realize that a machine is manipulating us to buy such a thing or vote for such a politician. It will be the end of democracy, predicts Harari. Worst of all, we’re contributing to that because we keep feeding Chatbots data.

Without knowing it, we are teaching things to that Matrix every day, educating it even better than we would educate our children.

When a child begins to speak, he hears minimal verbal stimuli in his family or social environment, perhaps 300 words? And with that verbal poverty, the toddler babbles and builds his language before school. Well, a ChatBot handles 1000 times more data than a child learning to speak. If you imagine that in the end, they are just texts, just put a voice to it, then ask the AI to add an avatar face and why not use a 3D printer and make an android? Well, they are already doing it right now. In other words, the endearing character Data from Star Trek will soon be among us. Like him, he will not be made of flesh and blood but plastic and silicon.

I insist. All that trail of information we are leaving on the Internet, even now, will be sooner or later captured by AI.

We are still concerned about providing our personal data every time we enter a website or social network, giving away not only our address and telephone number but critical information about our very identity and our preferences at all levels. And all of this, knowing that it has value, we give it for free, despite knowing that this digital gold is bought and sold for use by marketing companies and political parties to influence and manipulate us.

Well, now, it’s worse. Everything we write or say and display digitally ends up being sucked into these AI wits without us even knowing it and without getting paid for it. As the Wall Street Journal points out, we are facing a massive theft of intellectual property without any kind of compensation.

Users may never get paid, but sites and networks like Twitter have woken up and already charge ChatBots to access their data. The media, newspapers, radio, and TV stations distributed on the net also consider sending invoicing the owners of these machines that puncture their content and extract all the juice to further refine their linguistic models.

AI companies may be charged for using the information but reporting them for intellectual property theft is almost impossible. The legal problem is that Chatbots cannot be accused of plagiarism because they copy and generate language which turns out to be different from the copied source.

In other words, they do what the creators of cultural content have always done: collect and then transform. That process is the first step to learning and creating. In the XIX century, the musician Franz Listz traveled through the Hungarian countryside to listen to the folklore of the villages and compose his melodies by mimicking the same original music.

I do not intend to compare a machine to the geniuses of humanity. I am only trying to explain how ChatBots absorb and transform the information we create and publish freely, including me, because we want to be seen, heard, and get a high ranking in Google.

It is too late to prevent Artificial Intelligence from penetrating that gigantic mine of data, the Web. Blocking its way would be like shutting the barn door after the horse has escaped. But it would be fair that this big business of AI reverts in one way or another to the content creators while the privacy of all is guaranteed. 

This has not been written nor commented on with Artificial Intelligence… Oh maybe, yes?