Microsoft’s Safety Copilot brings AI to safety, and it’s making me nervous

Microsoft is bringing AI to safety, and out of the blue Skynet doesn’t appear so farfetched.

In a weblog submit, the corporate introduced Safety Copilot, a brand new device that’s purpose-built for safety professionals. Microsoft says that that is the “first safety product to allow defenders to maneuver on the pace and scale of AI. Safety Copilot combines this superior giant language mannequin (LLM) with a security-specific mannequin from Microsoft.”

Safety Copilot is constructed utilizing OpenAI’s GPT-4 generative AI, identical to the Microsoft 365 Copilot that the corporate introduced a few weeks in the past when it revealed its new AI instruments for Phrase, Excel, Powerpoint, and the remainder of its productiveness suite. This new copilot, nonetheless, makes use of a “security-specific mannequin to deploy expertise and queries that maximize the worth of the most recent giant language mannequin capabilities.”

When Safety Copilot receives a immediate from a safety skilled, it makes use of the total energy of the security-specific mannequin to deploy expertise and queries that maximize the worth of the most recent giant language mannequin capabilities. And that is distinctive to a safety use-case. Our cyber-trained mannequin provides a studying system to create and tune new expertise. Safety Copilot then will help catch what different approaches may miss and increase an analyst’s work. In a typical incident, this enhance interprets into beneficial properties within the high quality of detection, pace of response and talent to strengthen safety posture.

Microsoft says that Safety Copilot can have the next advantages for safety professionals:

  • Ongoing entry to the most superior OpenAI fashions to assist essentially the most demanding safety duties and functions
  • A security-specific mannequin that advantages from steady reinforcement, studying and person suggestions to fulfill the distinctive wants of safety professionals;
  • Visibility and evergreen menace intelligence powered by your group’s safety merchandise and the 65 trillion menace alerts Microsoft sees day-after-day to make sure that safety groups are working with the most recent data of attackers, their ways, methods, and procedures;
  • Integration with Microsoft’s end-to-end safety portfolio for a extremely environment friendly expertise that builds on the safety alerts;
  • A rising checklist of distinctive expertise and prompts that elevate the experience of safety groups and set the bar greater for what is feasible even underneath restricted sources.

I’m positively not a safety skilled, however the concept of bringing AI to safety instruments makes me extremely nervous. I’m positive my considerations aren’t legitimate, and precise specialists can clarify why that is necessary and mandatory, however I WATCHED TERMINATOR.