Global Security Forum

 View Only

An X-Force take on AI: What will it take to secure AI models? Breaking them.

  • 1.  An X-Force take on AI: What will it take to secure AI models? Breaking them.

    Posted Thu September 21, 2023 08:15 AM

    I was fortunate enough to have the opportunity write an opinion piece for The Hill on the security of artificial intelligence. As leaders in the AI space, I think it's important for everyone at IBM to understand how far we have come with regards to a security first mindset and the value of offensive security when it comes to emerging technologies. While the headlines are dominated with doom and gloom detailing the risks and ways to attack AI, I believe this is proof that we are actually making progress in cybersecurity. This is the first time in my career where we (security pros, policy makers, industry leaders, everyone!) are collectively thinking about the cybersecurity implications of a new technology BEFORE rolling it out.

    Every time someone learns how to exploit a model we learn a valuable lesson that will help prevent a future incident. Red team AI today so we don't have to IR AI tomorrow.

    Read the full piece here: https://thehill.com/opinion/cybersecurity/4192144-what-will-it-take-to-secure-ai-models-breaking-them/?sf181951054=1



    ------------------------------
    John Dwyer
    Head of Research, X-Force
    ------------------------------