Artificial Intelligence (AI) scares many people. The idea of creating a new intelligence that can think faster and better than humans, reprogram itself, multiply, and act beyond the machine that contains it, could potentially harm humanity more than all viruses combined. We need to think about this and protect ourselves.
At the beginning of 2023, I read “Prey” by Michael Crichton (yes, the same author of “Jurassic Park“!). Interestingly, this 2002 book, which is not about AI but about nanotechnology and nanomachines (or nanorobots) achieving independent life beyond human control, reminds us of the alarmist speeches that emerged everywhere following the explosion of generative AI at the end of 2022. Indeed, the media began to flood us with predictions ranging from the most optimistic to the most pessimistic about the future of our society.
In 2023, we no longer talk much about nanotechnology or the promise of nanomachines. Instead, we talk about the revolution linked to AI that hit us hard since the public release of ChatGPT. When OpenAI made ChatGPT publicly accessible and free in September 2022, humanity was propelled, somewhat forcefully, onto the fast track of AI.
But let’s not be mistaken; this doesn’t mean there was nothing concrete in this field before that date. Nor was it a secret. AI-based products or those relying on AI to function better have existed for a long time. For example, facial recognition has been present since the early versions of Google’s image management tool Picasa at the end of the 20th century, even before the shift to its Photos product.
For years, chatbots have appeared on support sites. These are small windows that look like a regular chat window, but instead of talking to a physical person, you talk to software. There are dozens of other AI-based technologies, or deep learning, or machine learning. The latter two are different and far below AI’s capabilities as they are primarily based on a training model.
In machine learning, algorithms are provided with data and answers. These algorithms then exploit this data, transform it into knowledge, and make good decisions based on this data. This is also known as data-driven strategy. Machine learning includes deep learning. This learning can adapt without receiving specific instructions on the subject. It combines statistical models with algorithms to analyze and establish inference patterns between the data submitted to it.
In deep learning, a machine is allowed to learn by itself but in a given subject, such as facial recognition. Instead of comparing pixels, deep learning enables the machine to recognize more abstract features than pixel values, like the shape of a face, mouth, hair, etc. The machine itself builds its set of characteristics. Still based on learning, identified data must be provided for the algorithms to deduce useful characteristics for future recognitions from unidentified data.
Finally, AI, which includes the previous two techniques, tries to reproduce or simulate not only human intelligence but also creativity. AI receives data and reacts to it but must also consider what it has done before. It must be able to analyze and adapt the effects produced by its previous actions. All this is done autonomously, without human intervention.
Thus, the various technologies briefly presented above are already found in commercial products or free OpenSource projects for several years. With the online release of ChatGPT in November 2022, nothing new under the sun for the IT community to which I belong. And yet, this event will fundamentally change the world. Why? Simply because OpenAI, with the publication of ChatGPT, offers the average internet user almost unlimited access to a tool of power they did not yet suspect.
For IT professionals like us, this technology represented only an evolution of something already known and used by companies needing to analyze myriad information to optimize their operations (such as insurance, banks, telecom operators, etc.). These techniques were already used in Big Data management, these gigantic volumes of data collected, for example, at every point of a national telecom network to predict problems before customers even notice them. But now, the average person is getting into AI, discovering fascinating tools (ChatGPT, Dall-E, etc.) and using them extensively, generally to boost their academic results, produce ten times more content and profit from it, make things that never happened seem real with entirely AI-generated photos, and more.
Recognizing that throughout history, humanity has shown a lot of recklessness, one could curse OpenAI for daring to open Pandora’s box of AI. This AI is presented by some as replacing many jobs held by humans. This AI is also presented as the nemesis of modern humanity (remember Skynet in the Terminator movie series!).
But let’s be careful.
Yes, ChatGPT and its peers have accelerated this movement. By offering access to their tools to the whole world and proposing paid subscriptions, OpenAI followed by its competitors have embarked on a rapid escalation of technological revolutions around AI, with generative AI being the most visible as it is intended for everyone. Each announcement from one is followed by an announcement from the other, arguing for more performance, more realism, more capabilities, etc.
But, and this is the important point in my eyes, it was done openly, under the sometimes amazed, sometimes offended, even frightened eyes of the whole world. What seemed like a great catastrophe had the merit of opening the eyes of our politicians who will legislate on the subject. The abusive use of AI tools by lazy students led to the implementation of techniques to detect texts, photos, and videos generated by AI. It is also thanks to the publicity around AI that we see the emergence of multiple countermeasures against AI.
So let’s not panic.
It is far too early to think about preparing a “survival kit” to face possible AI excesses. These tools are now part of our daily lives; let’s remain admiring but cautious towards them. Let’s use them, but let’s be careful not to make them our new deity.
