Is the protection of personal data at risk at the time of artificial intelligence?

And there are several billions of dollars at stake. Openai the creative company of ChatGPT informed that he is looking to gather up to 40,000 million dollars in an investment round; Goal estimates investing 65,000 million dollars this year; Amazon allocates 5 billion dollars in data centers; President Donald Trump revealed the Stargate project at a cost of 500 billion dollars for the United States; and the European Union announced the investment of 206,000 million dollars in AI.

In this race the data has a key role. Wester Zela, a specialist in artificial intelligence, indicated that the use of data has been training for traditional Machine Learning models. They were used by companies to understand the behavior of users, customers, consumers, and remained within that circle.

“The difference is that, with the emergence of these long language models, these LLMS, about 4 or 5 years ago that began to popularize, Not only are data sought within the company, but also sought data outside the company. That is, different data sources are accessed on the Internet, mainly ”said the conversation specialist with Commerce.

The problem with the use of personal data is the lack of transparency and, according to Zela, “it is a secret of the large technology companies, what data they use to train their AI models.”

This is also understood by Yamilet Serrano, academic director of Utec Postgraduate. “The difference is the scale. The AI ​​system are eager for data collection and Its inference is so little transparent that we have little control about what information it collects about us, why they use it and how we could modify or eliminate it ”he explained to this newspaper.

Mira: Chile will have its own artificial intelligence: Does Latamgpt do competition to Chatgpt and Deepseek?

More than one danger

The lack of transparency has been generating several risks, especially for users.

The report of Kaspersky Security Bulletin He pointed out that for this 2025 a greater sophistication of cyber attacks using artificial intelligence, through fraudulent messages in several languages ​​and Deepfakes, even using phishing, false websites or social engineering. And once again, the information takes importance.

Personal data have a value. Zela explained that when using a social network, application or platform on the Internet, we are leaving our information, even if the product is promoted as free, and “We don’t really know what is done with this information”. They can be used to training AI models, sold to other companies or stop in some black market. He compared the situation to what happens in Lima, when citizens data are then sold in premises of doubtful reputation, as in Wilson Avenue.

“Of course, they do not sell it in Wilson, but they are surely sold to other companies that pay millions, and these companies are worth what they are worth for the number of users, that is, for the amount of personal data they have. Once they are in the black market, because one is at the mercy of anyone, people who want to do business and people who want to extort and harm. We don’t know that “The specialist said.

Another question is the concentration of this information. Today, the main companies that work with artificial intelligence are OpenAi, Anthropic, Meta, Google, XAI, by the United States; Deepseek and Alibaba, in China; and Mistral, by France.

When consulted on whether this concentration can be a danger, Serrano considered that “It could be provided for the monopolization of information, which provides disproportionate power”.

Meanwhile, he explained that there are other risks from vulnerability to security gaps (a cyber attack can compromise data from millions of people), “Until the little transparency and control, since it is difficult for the end user to understand how their data is used and with whom they share it.”.

Look: The shape of the land internal nucleus is changing, according to scientific research

Regulate or not regulate AI?

This is the new debate, and that was pending at the Paris Summit. On the one hand, not regulate the development of AI could put people safety at risk, and on the other hand, it has been pointed out that regulation could compromise innovation.

For Serrano, “Today we all interact in one way or another with products that use AI, not regulate it would be to turn their backs ”, And for this, equity, transparency and security must be guaranteed. He considered it important to foster audits, establish independent supervision agencies and promote international standards for effective regulations.

Look: Valentine’s Day: Cybercriminals take the AI ​​to create false identities and recruit private information in appointment apps

But can innovation be stopped or delayed? Zela explained that this debate in technological development is not new, and there are usually two trends. While in Europe the regulation is chosen, in the United States the regulation is subsequent. This has led European companies to migrate to American land.

“Many of the companies you know now, and that they are very large, one thinks they are from the United States, but no. There are companies that were born in Europe, but that have had to go to the United States because there is a more favorable environment for growth ”he said, from his experience in both geographical areas.

In that sense, he indicated that there must be a midpoint, there must be a balance. If there is no data, there is no innovation, there is no training of AI models, it would not have existed ChatGPT. Thus, the challenge is in “Promote innovation and, on the other hand, also prevent citizens data from being committed”.

An issue that will remain pending, while artificial intelligence continues to progress. Recall that there are billions of dollars at stake.