CrowdStrike reports that AI is accelerating cyberattacks — helping criminals write more convincing phishing, find weaknesses faster and operate at greater scale. The tools that make defenders more productive are doing the same for attackers.
The Foresight take
In practice this means fewer tell-tale signs. AI-written phishing has fewer spelling mistakes and better context, and the time between a vulnerability being disclosed and exploited keeps shrinking. Speed is now the attacker’s advantage.
Defenders have to answer with speed too: faster patching, automated detection and response, and awareness training that reflects how good modern phishing has become. We use AI-assisted monitoring on the defensive side to help clients keep pace.
What AI actually changes
The significance is not a brand-new type of attack, but the collapse in the effort required to run old ones well. Generative tools let criminals produce fluent, context-aware phishing in any language, clone a familiar voice for a phone scam, and triage stolen data or scan for weaknesses far faster than before. The gap between a vulnerability being disclosed and being exploited — already shrinking — is narrowing further, and the old advice to “look for spelling mistakes” no longer holds. Volume and polish both go up at once.
Answering speed with speed
Defences have to match that tempo. That means behavioural detection and response that spots unusual activity rather than relying on known signatures, automated containment so a suspicious device is isolated in seconds, phishing-resistant multi-factor authentication, and a patching cadence measured in days rather than weeks. Staff awareness still matters, but it should now teach people that a convincing message is not proof of legitimacy — verification through a second channel is. We use AI-assisted monitoring on the defensive side so clients keep pace with attackers who are using the same technology.
And treat every unexpected request to move money, change bank details or hand over credentials as suspicious until verified through a separate channel, however polished the message looks.
Source: ITPro, 24 February 2026.