Duo: GitLab officially releases AI developer chat for Pro users
GitLab unlocks AI chat and adds privacy controls. With the announcement, the company also gives an insight into upcoming versions.
(Bild: PopTika / Shutterstock.com)
With the official release of the AI chat for development and testing, GitLab is introducing a new data protection function that will enable companies to control sensitive data. The aim is to reduce security and compliance risks when introducing AI.
After developers were able to test AI Chat Duo in the beta phase for a while, GitLab has now officially released it in version 16.11 for Premium and Ultimate customers. The associated functional innovation is the data control function: it protects sensitive data in the company and controls what the AI is allowed to read at project or group level. This is intended to prevent customer data from getting into the AI model training and being passed on in an uncontrolled manner.
The chat provides developers with answers in natural language on all aspects of coding and the platform. It suggests code, explains existing functions in the code or in the GitLab universe, rewrites and modernizes code and explains the DevOps functions of the GitLab platform. Developers also use Duo to create test models to check new sections for correct functionality and security before finally adding the new features to a project.
Duo is available in the GitLab online editor or as an extension for VS Code and JetBrains. Companies can add their own tools to it.
In the future: security checks and personalized models
In future versions, GitLab plans to release further AI chat features from the beta program for all Pro users: On the one hand, these are security checks that explain loopholes and offer solutions for them. On the other hand, the AI offers templates for teams to better organize merge requests and code reviews, for example.
As new AI functionalities, GitLab plans to enable companies to use personalized models tailored to their needs or to integrate their own models. Further information can be found on the GitLab blog.
(who)
