Why did artificial intelligence suddenly become hype?
Recently, in a conversation with Dr. Ugo Stocco, physician at Hospital Israelita Albert Einstein and fellow in the BioDesign – Innovation for Health course at Einstein, it was very clear the space that solutions using artificial intelligence (AI) are occupying in areas other than Information Technology (IT) .
But there are several doubts about its applications, efficiency, reliability, security, control and even legal and copyright issues, among others. To help understand these points, I went to talk to several specialists, and below is a summary of them.
With the launch of ChatGPT in December 2022, a new era in the evolutionary process of Artificial Intelligence is starting: the “Age of AI applied to business” . This era will be characterized by the invasion of solutions based on AI in operational, administrative, management and decision-making processes in companies.
And we are just at the beginning of this era, where many of the demands that today are strictly human will be supported or performed by AI algorithms.
A real example is Coca-Cola’s “Create Real Magic” project, in which a group of 30 “AI artists” are using ChatGPT (generative text) and Dall.e (generative images) to co-create promotional material for the Coca-Cola Company in an official company initiative.
Another example, this time in the more operational field, is the initiative by the investment bank Morgan Stanley in testing the GPT-4 technology for the creation of a virtual assistant that, based on a library with hundreds of thousands of articles in the sector, seeks to support the team of analysts with insights and relevant information faster and more efficiently.
But why only now?
In that same conversation with Dr. Ugo Stocco, the following question arose: “But what is the difference from everything that has ever existed in machine learning (machine learning) and AI to the current hype? Why does everyone now have to use the word “artificial intelligence” like it’s the world’s greatest discovery?”
In summary, until now we have not seen anything like the application power that ChatGPT demonstrated, nor in such a commercially accessible way. What we are witnessing in this era is the “unlocking” of the technological capacity that prevented artificial intelligence from evolving and the democratization of access to this technology.
There are three pillars that have evolved and that together are empowering AI applications in more complex scenarios based on huge amounts of data:
- Hardware – AI algorithms have always come up against the limitations of computational power. A lot of AI just wasn’t developed because it didn’t run on the computers of yesteryear. But nowadays we have enough computational power available to run more complex and cumbersome AI models.;
- Algorithms – In recent years we have evolved a lot in algorithms capable of dealing with complex situations (“non-exact” situations), creating a huge range of possibilities and applications.;
- Knowledge Bases – A knowledge base can be understood as the set of information/knowledge of an AI. In an analogy, if AI algorithms are computational representations of “how we think”, knowledge bases are digital versions of information, of “knowledge itself”. And with the high capacity to collect, store and access enormous amounts of information, we are experiencing an explosion in the creation of knowledge bases in all areas.
By putting all this together, “mass” applications of AI-based solutions began to appear, and so artificial intelligence is no longer something restricted to IT personnel and is starting to permeate other areas and interfere in processes, in society , in
decision-making, in power…
From this moment that we are experiencing AI is no longer something “hidden” and starting to have a more visible performance for the whole of society. In my point of view, we have reached a tipping point in AI, where some solutions already solve “human problems” in a way that is very close to or even better than us, and usually at much higher speeds and at low cost
But along with all this power and visibility, questions about its use are also appearing. This is because despite all the power that AI is demonstrating in dealing with uncertainties and complex (non-exact) situations, it is not perfect and, many times, it is not traceable, it is not controllable and it is not “understandable”, generating all this scenario of yearnings, insecurity and resistance to AI.
In conversation with several IT and business managers, the climate of caution and doubts is clear. Although everyone agrees that Artificial Intelligence will be a great game changer in the corporate world, many consider it “exciting and terrifying at the same time”, and its adoption is still a taboo, surrounded by concerns about control, predictability, compliance and security information, among others.