But the results of that mentality are real – and immediately. “Companies are moving very fast now,” says Rohds-Hare. “And this is the speed problem.”
New types of hackers for a new world
This rapid development has forced the security world to develop – but it has also been expanded to participate in it. While traditional pen-testers still bring valuable skills to the red team to the red team, the landscape is opening to the background and a wide range of subjects.
“There is a cycle of people that differ in different backgrounds,” says Sherets of Hacaron. “They cannot have a background of computer science. They cannot know anything about traditional web weaknesses, but they have some types of attendants with AI system.”
In many ways, the AI is less about breaking the security test code and is more about understanding the language – and, in detail, people. “Skills are getting good with natural language,” Sherrets says. It opens the door to examiners with generous art, communication, and even training in psychology – any person who is able to easily navigate the emotional area of conversation, which produce many weaknesses.
While the AI models do not feel anything themselves, they are trained on the huge trevings of human language – and reflect our feelings in ways that can be exploited. The best red teams have learned to bend in it, the crafting indicates that to get the system to break their rules, appeal to urgency, confusion, sympathy, or even manipulation to get a system.
But no matter what the background, sherets say, the required quality is still the same: “Hacker mentality … eagerness to break things and do them to do the work that others did not think.”
