According to GitLab’s annual global DevSecOps report, the use of AI in the development of new software continues to increase.

The developer platform GitLab has published the results of its annual survey. The global DevSecOps Report examines the global state of development, security and operation of online systems. AI is playing an increasingly important role in this.

Optimism is king

A quarter of the companies surveyed are currently using AI to drive human-driven software development. Another two-thirds plan to implement it. The vast majority of respondents, 83 percent, stress that incorporating AI into software development is essential to avoid falling behind.

The report seems to rate AI quite positively: no fewer than nine out of ten respondents have confidence in the day-to-day use of the technology. The integration of AI into the software development cycle is even experienced as very or extremely successful by some respondents.

In addition to coding, this technology is of course also used in many other areas of the technology world. The top 6 in terms of using AI were:

• Natural language chatbot documentation – 41 percent
• Automated tests – 41 percent
• Code change summaries – 39 percent
• Experiment tracking in machine learning models – 38 percent
• Suggestions on who can review code changes – 37 percent
• Summaries of comments on a topic – 37 percent

So generating code and proposing doesn’t seem to be the primary use of AI, coming in just behind the list above at 36 percent. According to this report, developers only spend a quarter of their time writing code, so that in itself is not surprising.

Issue

However, the questioning did not just trigger a hurrah. A third of those surveyed said they were very or extremely concerned about the use of AI in the daily software development cycle. Almost half of them are concerned about copyright protection of AI-generated code and 39 percent are concerned about possible vulnerabilities that code might contain. The contrast with the positive numbers suggests that many people still have very mixed feelings about AI.

What is emerging more and more frequently is the growing concern about whether AI will evolve to the point where it will take over entire professions. Almost sixty percent of the developers surveyed actually fear that AI will take over their role within the next five years.

40 percent of security professionals in this report also fear that AI will make their job more difficult. If you look at all respondents, this value drops to 29 percent.

Although AI is widespread, it is still a new technology. Not surprisingly, 81 percent of respondents say they need more training to use AI in their work. Six percent more than that, it is the job of organizations to train their employees enough so that they can smoothly use AI in their company.

A notable, but also logical, conclusion is that people who are already more experienced with AI are much more likely to associate the technology with increased productivity and faster cycles.

In May of this year, GitLab launched version 16 of its DevSecOps platform, with a big role for AI, of course. The company is also collaborating with Google Cloud to develop new generative AI capabilities.