Will Cyber Risk Kill Your GenAI Vibe?

A robotic figure taking a shower with a soapy sponge, expressing surprise as it glances at a shadow of a person holding a knife on the curtain.

Last updated: 14/08/25 | Published: 14/08/25

From productivity dreams to cyber risk nightmares

The impacts of GenAI on business IT are real. So far, the chatter has focused on developer productivity, starting with with GitHub Copilot's study claiming 55% efficiency gains on basic web development tasks through to Andrej Karpathy's introduction of "vibe coding" in February, 2025, in which a developer describes to a GenAI agent in natural language what program or app she or he wants written, and mostly lets the AI do the rest.

There's talk of 10x productivity gains and armies of coding robots managed by a single orchestrating architect.

Productivity, however, is just the first of many concerns for IT in business. Cyber risk is arguably a greater concern. Do you really want your business to be lightning fast at developing software that

leaks sensitive client or proprietary data, like Catholic Health + Serviceaide's leak of health data or McDonalds + Paradox.AI's easily stolen job-applicant data?
contains malicious, AI-agent installed code?
contains classic cyber risk NO-NOs like exposed credentials and easily guessed passwords?

Good GenAI vibes without cyber disaster

This blog post series highlights cyber risks to help increase the awareness of business managers, users and developers, thereby mitigating novel cyber risks arising with GenAI tooling and already existing cyber risks made worse by GenAI. Along the way we suggest mitigation strategies for for each GenAI cyber risk.

A guide for developers
A guide for product managers
A dedicated post about agentic systems

The responsibility for your businesses cyber security belongs with your business, not outsourced to some GenAI tool. Our posts will help you take charge of your GenAI cyber risks.