Microsoft sees uptick in AI attacks

Microsoft is warning of a growing threat from nation state cyberattacks and artificial intelligence (AI). According to the recently published Microsoft Digital Defense Report 2024, the cyber threat landscape continues to become more dangerous and complex, as threat actors employ increasingly sophisticated tactics, techniques, and tools that challenge even the world’s best cybersecurity defenders. “Nation state actors, activists, organised crime… all of these are putting in more resources behind these levels of attack, and more investment in terms of tools and changing tactics,” said Sarah Armstrong-Smith, chief security advisor at Microsoft in the EMEA region. “Each time we evolve, they evolve with it. But the real difference is the speed of these attacks. We have seen some threat actors able to get in and out in 15 minutes. Which is why rapid response, resilience and access to services is so critical to so many organisations,” she said, speaking at an Aon hosted cyber-risk webinar this week. According to Armstrong-Smith, Microsoft has seen an increase in cyberattacks involving AI, both against AI systems, and the use of generative AI by hackers to carry out cyberattacks and fraud. Over the next year, Microsoft expects a big rise in AI-automated fraud and election interference, and the use of deepfake impersonation to facilitate cyberattacks and fraud. “One of the things Microsoft has identified is not just deep fakes and disinformation, but how [threat actors] are experimenting with AI to increase the level of exploitation, reconnaissance, and really controlling the narrative in spear phishing campaigns and targeting specific organisations,” said Armstrong-Smith. Hackers are not only using social engineering to trick employees into giving away information, but also AI systems, according to Armstrong-Smith. “It’s clear that the threat actors are not only focusing on how to socially engineer and manipulate a human, but also how [to] socially engineer and manipulate a machine. We have seen an increase in specific attacks against AI itself, with regards to cross-prompt injections, trying to get AI to give up information or do things it would not normally be expected to do,” she said. According to the report, Microsoft is identifying around 600 million cyberattacks every day from nation states and cyber criminals. Microsoft now tracks more than 1,500 unique threat groups—including more than 600 nation-state threat actor groups, 300 cybercrime groups, 200 influence operations groups, and hundreds of others. The report warned of increasingly blurred lines between nation-state cyber activity and cybercrime. The US is the most targeted country for nation-state cyberattacks, along with Israel, Ukraine, Taiwan, South Korea, the UAE and UK among the most targeted countries. With wars in Ukraine and the Middle East, and ahead of the US election, Microsoft is seeing an increase in disinformation, deepfakes and electoral tampering. According to Armstrong-Smith, Microsoft has seen a “huge increase” in the use of proxy groups by nation states, which provides countries with a level of deniability and challenges for law enforcement, attribution and accountability. For example, Russia has outsourced some cyber espionage activities to criminal groups while Iranian and North Korean groups are targeting foreign companies for financial gain. Armstrong-Smith also warned that threat actors could be inspired by the CrowdStrike global outage earlier this year to carry out widespread disruptive attacks against supply chains. “It’s not all about cyberattacks. It’s been interesting to think back to the CrowdStrike incident in July, which was not a cyberattack; it was a change that went wrong. But it has really demonstrated the fragility of global supply chains and the requirement for digital resilience,” said Armstrong-Smith. “Arguably when thinking about some of these threat actors, when looking at the impact caused by the CrowdStrike incident, it could give them a playbook with regards to how to cause as much destruction across global supply chains, [and ask] what do I need to do to put that into fruition,” she said. From an insurance perspective, the threat from nation states and AI raises questions for cyber and digital resilience, according to Armstrong-Smith. “How prepared are organisations in terms of understanding and dealing with systemic risk, and the ripple effect of disruptive cyberattacks and large-scale outages, and the amplification of AI?... When thinking about the threat landscape, its cyber, disinformation, and AI combined,” she added.

Microsoft sees uptick in AI attacks

Want to read this article?

Read Next

Inflation and rising prices to drive Australian general insurance growth

Marsh expects European cyber rates to continue softening next year

Munich Re and ERGO target higher profit in 2025

Ransomware to target larger organisations next year, warns Moody’s

US commercial insurance market stable for buyers but sense of fragility remains

Inflation and rising prices to drive Australian general insurance growth

Marsh expects European cyber rates to continue softening next year

Munich Re and ERGO target higher profit in 2025

Ransomware to target larger organisations next year, warns Moody’s

US commercial insurance market stable for buyers but sense of fragility remains

Keep yourself informed

Sign up for Commercial Risk’s newsletters

Want to read this article?

Read Next

Inflation and rising prices to drive Australian general insurance growth

Marsh expects European cyber rates to continue softening next year

Munich Re and ERGO target higher profit in 2025

Ransomware to target larger organisations next year, warns Moody’s

US commercial insurance market stable for buyers but sense of fragility remains

Related Articles

Keep yourself informed

Sign up for Commercial Risk’s newsletters