Google says 2026 will be the year AI supercharges cybercrime
Security leaders are staring down a year of major change. In its Cybersecurity Forecast 2026, Google paints a picture of a threat landscape transformed by AI, supercharged cybercrime, and increasingly aggressive nation-state operations. Attackers are moving faster, scaling their operations with automation.
AI becomes central to both sides
By 2026, AI will be a normal part of everyday attack and defense activity. Adversaries are already using it to automate phishing, clone voices, and shape disinformation.
One of the fastest-growing threats is prompt injection, which manipulates AI systems to ignore safeguards and carry out hidden commands. As more companies deploy LLMs inside business processes, these attacks are becoming easier to launch and harder to detect.
AI is also changing social engineering. Groups such as ShinyHunters have used generated voices and realistic phishing to trick people instead of bypassing technology. Voice cloning is now cheap and convincing enough to impersonate executives or IT staff during vishing calls.
The report notes a growing reliance on AI agents, systems that act on their own to complete tasks. These agents will need distinct digital identities and strict access controls. Security programs built for human users will not be enough. Identity management will have to account for AI-driven decision making and temporary task-based privileges.
AI is also reshaping security operations. Analysts will soon direct AI tools rather than manually sort through alerts. Instead of reviewing logs, they will examine case summaries and confirm automated containment steps. This shift enables faster response but also brings new oversight challenges.
“While adversaries are certainly trying to use mainstream AI platforms, guardrails have driven many to models available in the criminal underground. Those tools are unrestricted, and can offer a significant advantage to the less advanced. There are several of these available now, and we expect they will lower the barrier to entry for many criminals,” said Billy Leonard, tech lead, Google Threat Intelligence Group.
A related concern is the rise of shadow agents. Employees may use unapproved AI tools to handle work, often without realizing the data risks. Banning these tools pushes the problem underground. The report recommends guardrails, monitoring, and governance for how AI systems are used internally.
Cybercrime keeps expanding
Ransomware and data theft remain the most disruptive threats worldwide. Combined attacks that encrypt systems, steal data, and pressure victims through public leaks continue to spread.
In the first quarter of 2025, more than 2,300 victims were named on leak sites, the highest number since tracking began in 2020. Attackers are exploiting software supply chains and zero-day vulnerabilities to reach hundreds of targets at once.
Social engineering remains a common entry point. Voice phishing and tailored messages still bypass MFA and other defenses. Extortion schemes are evolving beyond stolen data to include threats that halt operations or expose executives.
As more financial activity moves to blockchain platforms, attackers are using the same systems to hide their tracks and move stolen assets. Investigators now need to read smart contracts, trace wallets, and connect transactions across public ledgers. Blockchain transparency works in two directions. It helps criminals hide from takedowns, but it also leaves permanent records that can later be used for attribution.
With stronger endpoint defenses, adversaries are shifting toward virtualization platforms. By targeting hypervisors that host virtual machines, attackers can disable hundreds of workloads in hours. The report advises direct investment in securing this infrastructure, not just the applications that depend on it.
Industrial environments remain targets too. Criminals are striking enterprise software that supports operational technology, forcing quick ransom payments when production stops.
Nation-state activity expands
Cyber operations linked to governments will continue to grow in 2026, each driven by its own objectives.
Russia is expected to shift from short-term wartime operations in Ukraine toward longer-term global objectives. Information campaigns and hacktivist groups will continue to focus on Europe and North America, including election interference and infrastructure disruption.
China is likely to remain the most active state actor. Its operations emphasize espionage and stealth, targeting third-party service providers and edge devices that often lack monitoring. The semiconductor industry is a key focus as competition for AI technology intensifies.
Iran will keep blending espionage, disruption, and influence operations tied to regional conflicts. Propaganda and fake news sites are expected to use AI-generated content to amplify pro-Iran narratives.
North Korea will stay focused on cryptocurrency theft and intelligence gathering. In 2025, North Korean groups were linked to a theft of about $1.5 billion, and those operations are expected to continue. Some of the country’s IT workers are taking remote jobs abroad to gain access to corporate systems and digital wallets.