AI: every activity needs norms and rules

All activities in life are subject to norms or rules. It is one of the characteristics of civilized societies. It is evident that advances, as happened back in the day with the internet, begin by breaking some ways of doing or working, but over time they become normalized, although there are always those who find a way to go down the path of illegality or know how to take advantage of the lack of rules. We must be careful to reduce this aspect as much as possible.

Each advance is implemented exponentially faster than the previous one. The radio took longer to become massive than television, and the latter longer than the internet. And they have all ended up being regulated in one way or another. Artificial intelligence (AI), which permeates all activities and even the day-to-day life of citizens, following the pattern, has become generalized even more quickly. ChatGPT, from Open AI, was launched in 2022 and has not stopped growing since then.

And the very people who have set this technology in motion and launched tools to use it, from OpenAI to Anthropic, have raised an outcry and asked for regulation, supervision, and to reduce the growth rate of a technology with the capacity to end up dominating humans. But the messages have diverse objectives, ranging from possible risks to geopolitical struggle or the economy. There are dangers, as in all areas of life, but in this case, with a greater destructive potential. Today, everything depends on the hand of man, but systems that self-learn, if not well directed, can end up doing harm instead of good. On the other hand, there are also interests.

On the part of US President Donald Trump, those of surpassing China in this race, without regulations or obstacles. Beijing, for its part, wants facilities to expand its technology and take advantage of the occasion to strengthen itself as a power, as it did with 5G through Huawei. The big US tech companies want there to be barriers to entry (or even walls) for new competitors, so that it is much more difficult for them to break their oligopoly on a global scale.

And there are also economic interests. With stratospheric valuations based on high expectations, specialized companies have looked towards the stock market. But the context has forced OpenAI to postpone its entry into the market. In addition to its valuation decreasing because there are competitors that surpass it (every few months one model surpasses the other and the amortization of investments becomes impossible), there is the fear that potential misuse of the technology will cause a disproportionate stock market punishment. These companies are also aware of the need for resources for data centers, in competition with state debt, which are forced to raise the profitability of their debt to make it more attractive.

Functioning in society requires norms and rules, and also a new technology like AI, but, furthermore, these norms should be global, positive, and block misuse as much as possible, as happens (or is attempted to happen) with nuclear energy or the internet. It cannot be allowed that the regulation of AI is made by an oligopoly in the USA, under the perspective of its interests; nor in China under theirs. The European Union (EU) has already done its part. Now it would be necessary to negotiate a global one, as has been attempted with the taxation of multinationals through the OECD.