GenAI could make software much less buggy by 2028 with better unit tests

If you’ve ever studied computer science at university or taken an online course in programming, there’s a very good chance that you’ve come across the concept of unit testing, where you create tests to check that small portions of your code are doing what you intend.

This aspect of coding, according to the analyst IDC, will be largely taken over by generative AI (GenAI) by 2028 with a massive 70% of such tests being created by AI. Some programmers find unit testing a little bit confusing too, so having AI write better tests should improve the overall quality of code shipped to end users.

Discussing the adoption of GenAI in Asia to create tests, Dhiraj Badgujar, Senior Research Manager for Digital Innovations, xOps and Developer strategies, IDC Asia/Pacific, said:

“The Asia/Pacific market for generative AI assisted software development, including code generation, user interface, testing, and other use cases, is rapidly growing. China, India, and Japan are investing significantly on this to enhance software testing capabilities. The IT sector's robustness and quality assurance focus are driving market progress”

With GenAI writing better and more comprehensive unit tests, it will free up the time of developers who can instead focus on creating new features for the software they are working on. As mentioned, these comprehensive unit tests should also catch more faults with the program so that software will be less buggy.

The idea that we will get safer software thanks to AI is actually pretty interesting when we consider the increasing popularity of Rust which comes with memory safety built in.

One of the reasons there are lots of exploits out there is due to memory issues in software created with low-level programming languages. If more software is made with Rust and AI is being used to make better unit tests then we could potentially see exceptionally solid software being churned out by the end of the decade.

Source: IDC

Source