The call comes in the wake of a notable increase in requests to modify AI-generated software applications.
Bootham-based I-Finity says many organizations are turning to AI tools to quickly build apps, platforms and internal tools, only to run into problems when trying to deploy or scale them.
Industry and academic studies support these findings, showing that code generated by AI can contain measurable security vulnerabilities and logic flaws.
Recommended books:
As more organizations rush to ship products built with AI, the cost difference between prototype software and production-ready software is becoming increasingly large.
According to I-Finity, AI-generated code is highly effective for creating quick demos, proofs of concept, and early prototypes.
However, enterprises are realizing that many AI-first builds are plagued by bugs, performance issues, scalability, and complex compliance requirements when put into real-world use.
Russ Huntington, Chief Technology Officer at I-Finity, said: “AI is great for exploring ideas quickly, but more and more companies are finding that apps built with AI are not designed to scale or meet compliance requirements.”
So-called “vibe coding,” which uses AI tools to generate code, is becoming increasingly popular.
Searches for the term have increased from virtually zero two years ago to more than 33,000 searches per month.
“Vibe Coding helps lower the barrier to entry, reduce upfront costs, and allow teams to explore ideas without committing to a full development cycle. However, it often fails when initial builds go further than the design.
“And if issues around GDPR and secure data storage are not addressed early, instant-fix AI apps could quickly become a significant risk.”
Vibe coding is also used by professionals. 84% of developers use AI coding tools in their workflows, and these tools are responsible for approximately 41% of all code written in production.
And this led to problems with almost half of developers reporting quality issues and incorrect output from AI coding tools.
According to a study by Sonar, 72% of developers use AI tools every day, but less than half review their code on an ongoing basis.
An independent study from Cornell University found that code generated by AI can contain measurable security vulnerabilities and logic issues. The quality of the final code varies widely depending on the tool.
I-Finity advises companies to treat AI as a starting point, rather than a replacement for a proper engineering or custom software development team.
