Cybersecurity and AI: New Threats and Practical Defences

Artificial intelligence is changing both sides of cybersecurity. Attackers use it to create more convincing messages at scale, while defenders use it to detect suspicious activity faster. Businesses need to update their habits accordingly.

New and improved threats

• Polished phishing: emails and messages with perfect grammar, personalised details and local language.
• Voice and video impersonation: cloned voices and fake video calls used to request urgent payments.
• Faster reconnaissance: automated scanning for vulnerable websites and exposed credentials.
• Risks from AI tools: staff sharing sensitive data with public chatbots, or trusting insecure AI-generated code.

Defences that still work

• Verify requests out of band: confirm payment or password changes with a phone call to a known number.
• Use multi-factor authentication: ideally with passkeys or authenticator apps rather than SMS.
• Agree a code word: for urgent financial requests within your team.
• Keep systems patched: most attacks still exploit known, fixable weaknesses.
• Back up offline: so ransomware cannot destroy every copy.

Use AI on your side

Email filtering, anomaly detection and endpoint protection increasingly use machine learning. Choose reputable providers and keep human oversight.

Review AI-generated code

Treat suggestions from coding assistants like code from a junior developer: review it, test it and scan for vulnerabilities before it reaches production.

Train regularly

Short, realistic training sessions help staff spot new tricks. Encourage a culture where it is safe to report a mistake quickly.