“We May No Longer Have a Critical Sense of What We Are Receiving”: Generative AI — Threat or Ally?

Experts from Cisco, Sport Lisboa e Benfica, MediaLivre, and Vinci Energies Portugal came together at the IT Security Conference 2025 to talk about how artificial intelligence is changing digital security in different industries.

The main idea that came up was that the line between innovation and danger is getting harder to see, as AI is both a tool for protection and a way for threats to grow.

Rui Antunes, who works in cybersecurity at Cisco, said AI is helping both good and bad people.

“Emails made by AI are now so realistic that we often don’t realize they’re not real,” he said. He also talked about the risks of data leaks and how hard it is to predict what generative models will do. “If a company doesn’t really understand how its app works, how can it tell if its AI is being used wrongly?” he asked. He said this uncertainty can cause big problems for a company's reputation and money.

Paulo Martins, who runs IT and operations at Sport Lisboa e Benfica, said there's a bigger risk of fake information now.

“Even the content we make in digital and broadcast formats can be changed a lot,” he said. “We might not be able to tell what’s real anymore.” Luís Amorim, from MediaLivre, added that fake videos and deepfakes are making it hard to trust news and could harm democracy. “Our business is in danger,” he said.

In the energy industry, Miguel Gonçalves, the CISO at Vinci Energies Portugal, talked about something called “data poisoning” in important systems.

“It can damage more than just data-it can also hurt people’s safety,” he said. He believes in good leadership, ongoing training, and smart ways to watch for threats.

Even though there are issues, the group agreed that AI can be useful for protection.

Cisco uses AI to keep track of 550 billion events every day and can find problems quickly. For Benfica, AI helps fans enjoy the game more and makes the team run more smoothly. MediaLivre uses AI to give readers better personalized content, but Amorim said, “People will always be the ones making the content.”

In the end, Gonçalves said the real question isn’t whether AI is a threat or a helper—it’s whether companies are ready to understand, control, and use it in a smart, ethical way before it gets out of hand.