The AI authorship crisis is becoming a cyber-security problem

The debate around generative AI and authorship is no longer confined to publishing and the creative industries, it is becoming a cyber-security issue too.

 

A recent BBC report highlighted growing concerns from writers over how AI systems are trained and how easily human work can be replicated. In cyber-security, the same problem is emerging around communication, identity and trust.

 

For years, employees were trained to spot phishing through poor grammar, unusual wording or suspicious tone. Those signals are disappearing. Generative AI can now produce emails and messages that closely imitate internal communication styles, making impersonation attacks significantly more convincing.

 

The UK National Cyber Security Centre (NCSC) has warned that AI is already increasing both the scale and effectiveness of phishing and social engineering campaigns.

 

At the same time, attackers are combining AI-generated text with voice cloning and synthetic identities. Europol’s latest Serious and Organised Crime Threat Assessment describes AI as a growing accelerator for fraud, impersonation and organised cybercrime.

 

This is creating a broader authenticity problem for organisations. Identity security is no longer only about passwords and authentication. Increasingly, businesses must verify whether the person communicating with employees, suppliers or customers is real at all.

 

That is forcing organisations to rethink how trust operates internally. Verification is becoming more important than instinct. Multi-factor authentication (MFA), phishing-resistant login methods and secondary approval processes are becoming essential in environments where AI-generated communication can appear legitimate.

 

The NCSC has also warned about the need to preserve the integrity of digital information as generative AI becomes more embedded across communication systems. Its guidance on generative AI and information integrity reflects a growing concern that authenticity itself is becoming harder to verify.

 

The AI authorship debate may have started in publishing, but the underlying issue is increasingly relevant to cyber-security too: when machines can convincingly imitate humans, trust becomes a security challenge.