Cybersecurity Ventures predicts that by 2025, cybercrime will price the world $10.5 trillion yearly, an enormous enhance from $3 trillion in 2015, with a lot of the rise attributed to using superior applied sciences like LLMs.
Take attribution – many have used an LLM to put in writing “within the voice of”- however attribution is that rather more troublesome in an AI world, as a result of menace actors can mimic the strategies, feedback, instruments, and TTPs. False flag occasions turn into extra prevalent, such because the assault on U.S. service members’ wives.
LLMs are accelerating the arms race between defenders and menace actors, reducing the barrier to entry, and permitting assaults to be extra complicated, extra insidious, and extra adaptive.
Extra broadly, adversarial AI is coming on your Apps, writes Dan Shugrue, a safety product advertising chief at Digital.ai, in a TechRadar Professional article.
Learn the Full Story
