Presented by Microsoft
Cybercriminals operate with greater agility than corporations, unencumbered by the need for executive approvals, testing, and regulatory checks. This positions many organizations at a disadvantage, akin to showing up to a gunfight with a butter knife. However, generative AI could significantly alter this balance, as discussed by Kelly Bissell, Deputy CISO and Corporate VP at Microsoft, during the AI Impact Tour in Atlanta.
“What I hope we can achieve as a community is the swift and safe adoption of AI, allowing us to gain the upper hand against attackers and avoid being left behind,” Bissell stated.
Supercharging Security with Generative AI
Security is one of the most compelling applications of practical AI. Early users of Microsoft Copilot for Security have observed its effectiveness in incident response—particularly in mitigating attacks and enhancing the Security Operations Center (SOC) —all while dramatically cutting costs.
Informed by vast data and threat intelligence, including over 78 trillion security signals processed daily, Microsoft Copilot leverages large language models to provide tailored insights and actionable guidance. This technology simplifies the detection of subtle attacks and significantly reduces response times, enabling security teams to stay one step ahead of cybercriminals. Its user-friendly interface allows analysts to pose natural language queries and receive actionable responses.
These tools equip new, junior SOC analysts with the skills needed to perform effectively from day one, decreasing the time and costs associated with onboarding. The time required for scripting or incident analysis drops from days or hours to just minutes.
“I have a pharmaceutical company that expects to save $50 million annually within their security team while enhancing their security,” Bissell revealed. “This dual benefit—improved security and reduced costs—is a common theme among more than half of the companies we work with.”
Early adopters have also recognized the importance of training and transparency. AI tools require thoughtful use, and Bissell emphasizes that they’ve integrated logging features to allow clients to monitor AI operations. Furthermore, a red team tool has been added for verifying results, fostering a “trust-but-verify” approach. This added transparency has enhanced both the platform’s effectiveness and client confidence in its capabilities.
Generative AI Security in Action
Bissell highlighted the pharmaceutical sector as a prime example of how tools like Copilot for Security can create a wide range of effective applications. Initial skepticism has paved the way for greater exploration of AI capabilities.
One significant advantage is reducing fraud during clinical trials. Malicious actors have financial interests in whether a drug passes FDA inspections, posing risks such as falsified results and insider trading. By applying AI-driven security to operational technology in manufacturing plants and labs, companies can both monitor equipment for signs of failure and detect potential tampering. This integrated approach is also enhancing drug and polymer research.
“They can produce better products and accelerate the go-to-market process for drugs, which is worth billions of dollars,” Bissell emphasized. “With a patent lasting only 10 years, being able to enter the market sooner allows companies to maintain their market share before it becomes publicly available.”
Transforming the SOC could be the most impactful use case, especially as cybercriminals begin leveraging generative AI without the regulatory hurdles organizations face.
“We’ve noticed a significant rise in the use of open-source AI among attackers who can now build models more readily,” Bissell explained. “You don’t need advanced hacking skills to create malware anymore. Just as developers can rapidly produce code through platforms like GitHub, so too can cybercriminals. The number of attackers is set to rise dramatically. This race has only just begun.”
Ranking
