Artificial intelligence will define 2025, a year marked by robotics and AI agents

What is expected in 2025? Well, we are facing a promising year in technological innovation. Since the launch of ChatGPT, the range of AI development companies has expanded.

OpenAI, creator of this chatbot, continues to lead companies worth $86 billion, according to last year’s ForbesAI 50 list. Anthropic or Databricks are next in importance. But in this race for AI there are also more traditional manufacturers such as Google, Meta, Microsoft, IBM or xAI from Elon Musk that drive supply.

For Diego Castronuovo, head of AI at N5, there are two keys to understanding the IA en is 2025. On the one hand, consider that “There is a ‘terrible’ investment in infrastructure and research”which marks that AI will be the biggest technology in the coming years worldwide.

“The key factor is the search for artificial intelligence In general, what is known as AGI, which is neither more nor less than an autonomous system that is capable of developing intellectual tasks like a human being, equal to or better than a human being.he indicated to The Commerce the software company specialist.

This AGI is one of the best-known challenges of OpenAIor at least for its CEO Sam Altman. For this, the company has its o3 system. However, there are still voices skeptical of this possibility.

Another key factor is the role of AI agents, or autonomous systems that can perform tasks with less human intervention. Castronuovo points out that these are different models from those known as ChatGPT why “is dedicated exclusively to taking a task, analyzing it, dividing it into steps and executing it”.

Wester Zela, a specialist in IA and dean of engineering courses at the Scientific University of the South. In addition, AI agents make it possible to reduce hallucinations (errors) that an AI system may have.

“Now the issue of AI agents is taking center stage. Basically they are new ways to make the models generally have fewer hallucinations. “It is a new technique to be able to, let’s say, have better answers.”Zela indicated. This is a division of labor among agents specialized in different tasks.

But another important role in 2025 will be robotics plus artificial intelligence. Models like Atlas, Optimus and Ameca have surprised in 2024 due to their human characteristics. For the teacher, these functions will be seen more in 2025: “bring artificial intelligence to the physical world, through robots, through different devices”. Obviously, this will have an effect on daily and work activities. It’s about moving from software to hardware.

Will AI no longer be free?

Although the supply of artificial intelligence models increased in 2024, their increase in price has also been notable. They started as free models but now they are betting on a subscription business model. This is the case of OpenAI, which can charge up to US$200 per month in its Pro version of ChatGPT.

Zela indicated that it is the model freemiumin which a service is free and, based on new functionalities, it is paid. He explained that keeping an AI model running costs millions of dollars and “You have to make money in the same way as in many other ventures”.

The detail would be in the competition, since “the idea of ​​subscribing to a license for any tool has to have some advantage”.

Opposite payment models there are also OpenSource, or open source, models. One of these cases is the Llama de Meta model, which any user can access.

Castronuovo considered that in any case “We are always assured of a level of free access on the Open Source side and by models of large multinationals that generate profitability”.

“We believe that the free model, with access to the general public, is definitely available and will continue to be, and we are going to have this other issue related more to the field of research and professional, which points a little more to the issue of charging.”he indicated. It is a situation that happens with N5, a software company that creates AI solutions but for the financial field.

MIRA: CES 2025: what will be the technological innovations that will mark the beginning of the year?

El factor Donald Trump

2025 also begins with the return of Donald Trump to the presidency of USA. But he doesn’t do it alone. With him is the presence of Elon Musk and the support of Sam Altman for the development of AI.

In a recent report by The CommerceMaité Vizcarra, an expert in technological innovation, pointed out that in this new management “lost time will be made up for” and is expected to support deregulation, since “during the time of [Joe] Biden, there has been a lot of suspicion in the American government and there has been a lot of discussion about whether to regulate or not regulate.”

The new position of the United States could clash with that of Europe, which has strong regulation of AI: in August 2024 the regulation came into effect. Artificial Intelligence Lawthe first in the world to guarantee security and respect for fundamental rights.

One of the possible effects is increased investment in AI, but also polarization with China. “It is a dynamic that can generate two poles: an American pole and a pole in China, generating artificial intelligence with a deregulated market on the United States side. It is a little more complex to understand how they are going to generate the security of these models”Castronuovo indicated. AI models could be used for inappropriate and malicious purposes if they fall into the wrong hands.

MIRA: Financial attacks on smartphones, sophisticated scams with AI and crimeware in Chinese: know the 10 new threats for 2025

Cyberattacks + AI

And yes, data security is also a very delicate issue to deal with, especially if money is involved. 2024 has been marked by cyber attacks on users and financial entities. The Interbank case had a strong impact last year in Peru. It was in the month of October when a malicious actor claimed to have accessed sensitive data of clients of the banking entity, which confirmed the attack with the aim of extorting them.

Can artificial intelligence increase security risks in 2025?

Castronuovo makes two key points in this paper. On the one hand, the use of artificial intelligence for defense, that is, to predict behavioral patterns that can warn of fraud risks. “By 2025, it is expected that these solutions related to early detection will begin to grow more”he indicated. But the real danger is in the AI ​​that generates attacks.

The financial security specialist pointed out that this is a challenge for the future, when AI is not designed for defense but for attack. “One thing is human beings sitting behind a machine hacking systems, and another thing is artificial intelligence trying to hack systems. “They are very different things.”he indicated.

MIRA: What is the origin of ‘Chill Guy’, the meme that is sweeping social networks, and what does it mean?

On the other hand, according to the annual report Kaspersky Security Bulletincyberattack sophistication is expected using AI to hone fraudulent messages in multiple languages ​​and deepfakes with highly credible quality, in advertisements, phishing emails and fake websites.

AI in Latin America and Peru

Artificial intelligence also has an effect in the region.

If the specialists consulted also agree on something, it is that Brazil and Chile lead the development of AI in the region. Mexico and Argentina can also be added to them, which has recently taken steps towards AI.

In general, for the N5 specialist, Latin America has a particularity: inequality in technological infrastructure in certain countries.

“There is significant inequality, as perhaps there is not so much in Europe. These levels of inequality present some challenges for the growth and adoption of this technology. I think that hand in hand with that, prioritizing the training of talent is also another key point. It has been scarce, but not because it does not exist but because a lot of talent has decided to work for developed countries”explains Castronuovo about our region.

MIRA: What is expected in 2025 from artificial intelligence, the advance that marked a before and after in the history of technology

To overcome this situation, mutual collaboration between countries is proposed. Furthermore, private companies will have to “start finding business cases to apply artificial intelligence that ensures profitability”.

And what happens in Peru?

The situation is a little more complicated. Professor Wester Zela points out that “beyond having a Law, beyond having a strategy that remained on paper, or individual initiatives of institutions, “There is no leadership that really wants to bet on artificial intelligence”.

Our country has legislation. In 2023, Law No. 31814 was enacted, which promotes the use of AI for the economic and social development of the country. However, in the words of Dr. Zela, “concrete actions are not being seen”.

MIRA: Apple denies that it is going to launch its own search engine in competition with Google

In that sense, he proposed that the development of AI in the country should go through the creation of a National Center for Artificial Intelligence, a High Performance Computing Center and the use of updated data in open format so that researchers can use that information. information and create your models.

“We would necessarily have to invest as a country. For example, Chile has a very powerful computing center, a high-performance computing center that provides services to academia, that provides services to the public sector and that provides services to the private sector, especially small and medium-sized businesses.he explained.

Now, although Chile may be more advanced in the management of AI, it is worth noting that Peru has the High Computational Performance Center where the MANATI Supercomputer is present. The website of the Ministry of the Environment points out that it was installed in Iquitos in 2017 as a computer laboratory to support scientific and technological research.

In the case of universities, there is also a pending task: the training of professionals specialized in AI. “With one university, with two or three, it is difficult. “There has to be a decision at the government level.”said Zela, who promoted the AI ​​career at the National University of Engineering and is now considering doing the same at the Scientific University of the South.

And this is how artificial intelligence begins in 2025.