Mexico City, Mexico – Artificial intelligence is making attacks on the cyber systems of companies and agencies more accurate and faster. Part of the problem is the use of artificial intelligence search engines for issues that have to do with companies, but over which companies have no control or visibility.
This is one of the issues on the table at the Microsoft AI Tour 2026, where Vasu Jakkal – corporate vice president of security for Microsoft – said that the technology company is looking closely at how to address the security of companies, particularly with artificial intelligence agent programs, or systems that monitor the security of a company.
Part of the problem is that both sides of the coin have not only these agents, but access to artificial intelligence (AI).
The company’s projection is that there are already around 1.3 billion AI agents. The cybersecurity expert spoke about the different challenges that AI represents for companies and, as has happened in Puerto Rico, for the government, which has faced cyberattacks that put citizens’ sensitive information at risk.
“What we know is that 80% of Fortune 500 companies have agents. But only 49% have security controls and 29% said employees are using artificial intelligence tools that they have no visibility over,” he said. “That abyss caused by the lack of visibility in the use of these tools versus what outside agents can see is a lot.”
Among the problems he identified with cybersecurity, he mentioned how quickly it is moving and the theft of sensitive data.
“The other problem we see is the issue of regulations. That the artificial intelligence agents in the agencies comply with the regulations,” he said. “The solution is to have all-encompassing security.”
Jakkal could not specify how much money in losses AI-assisted attacks mean, since these are expenses that exceed billions of dollars. He did say that the possibility of attacks for a company is 7,000 per minute.
“The time an employee or customer opens an email phishing At the time that this cyberattack accesses sensitive information, it is seven minutes. With artificial intelligence, that time can be reduced and the phishing It is very difficult to catch and now we see that they are more efficient,” he said.