Why is there an avalanche of alerts and fear of artificial intelligence?

BarcelonaIn recent days, even the main representatives of artificial intelligence (AI), from Open AI to Anthropic, Space X or Google, have warned about the risks of a technology that, they warn, may end up dominating humans. "There is a genuine concern about the models that are becoming common and have not been tested much and can be used to commit atrocities," states Esteve Almirall, Esade professor specializing in data, analytics, technology and AI. In this context, once Open AI, the creator of ChatGPT, has decided to delay its stock market launch to take a kind of pause in growth, some see the fear of a stratospheric punishment in the market if any of the existing tools is involved in some wrongdoing. Some of these companies have valuations that exceed the gross domestic product (GDP) of countries like Germany (more than five trillion) and a negative reaction would mean an almost unprecedented collapse given that there have never been companies with these prices, warns Josep Soler, founder and executive advisor of EFPA Spain (Spanish Association of Financial Advisors and Planners). "The enormous valuations that have been made of these companies would collapse in the event that trust in them was lost," he adds.

The aspect of the risks

British Anthropic engineer Jacob Coxon, who had also worked for OpenAI, became popular a few days ago when he submitted his resignation while warning that both companies "are not acting responsibly" and are "playing with our lives". In a thread on his X account, he warned that the people behind the exponential development of AI "believe that it could kill us all by the end of this decade". It did not take long for the main representatives of the companies to join the warnings, among them, the CEO of Anthropic. Dario Amodei asked in a public letter that the pace at which AI model capabilities are being improved be reduced. According to him, technology is currently moving towards a dynamic of "recursive self-improvement" that threatens to cause cybersecurity catastrophes on a global scale.

Although AI currently needs humans to direct it for good or bad purposes, it is not ruled out that in the future it could make decisions and act autonomously. Or that it could be used perversely. Anthropic, for example, discovered that the Houthis in Yemen used its tool Claude – paying with a credit card, like any other user – to direct their missiles against drones. And there are cheap and very powerful tools that can be included on a PC of 1,000 euros like Qwen, from the Chinese giant Alibaba, with which one can enter health or financial systems with vulnerabilities or entry points or make other perverse uses of them, warns Almirall. There is concern about these open, very powerful models, which any drug trafficker, terrorist, or extortionist can set up on their computer and which are difficult to find or stop, he adds.

The geopolitical aspect

The representatives of the AI oligopoly in the US are calling for it to be regulated. This is something that the more left-leaning wing of the Democratic Party, led by Bernie Sanders, also supports. And there is also a call for the creation of a kind of international body to oversee its activity and provide a kind of quality seal. In July, 29 countries signed an agreement in Shanghai to create the World Artificial Intelligence Cooperation Organization, an intergovernmental organization intended to promote international cooperation and AI governance. The President of the United States, Donald Trump, opposes it and prioritizes blocking China's path in the AI race and, therefore, having large US tech companies continue their race without considering other issues, such as with the race to dominate space. Beijing, on the other hand, is interested in such an initiative, which would allow it to operate even in the US, states Almirall. The Chinese Minister of State Security, Chen Yixin, has called for the creation of a "safety barrier" for AI and has situated it as one of the main fields of global technological competition. At the same time, he has warned of the risks involved in its accelerated development.

The Asian power, which is forging a technological silk road that passes through Africa and South America, sees in AI another opportunity to consolidate itself as a power, as happened with 5G and the role of Huawei. The leaders of OpenAI or Anthropic, for their part, would like there to be barriers to entry in the business, not only for the Chinese, but for other companies and initiatives from the USA itself to maintain their market dominance, experts state. It would be a matter of limiting the competitiveness of new entrants and turning the AI business into an oligopoly.

The money side

The AI race is intensive in the consumption of financial resources. Almirall states that "this hyper-competition significantly diminishes corporate profits, and it must be paid for. In fact, models have to be created every six months that no one can amortize." In addition to the stock market route, which Open AI has already delayed until next year, the big tech companies, aware that gigantic resources are needed, are leading fundraising in the markets, with more than 600 billion dollars, through debt issuances because they do not want to run out of cash. This, in fact, has not been the only factor, but it is one of those that have contributed to raising the profitability that investors have demanded in recent times on state bonds, which have to pay more to raise financing and cause the issuances of big tech companies to lose their appeal. "These companies, with unprecedented valuations based on almost unlimited expectations, are the ones that are dominating the capital markets," warns Soler. These stocks are the ones that, in the event of negative news or news that affects their reputation, could trigger a stock market earthquake.