The AI coding assistant supplier Tabnine immediately introduced a brand new characteristic, Code Provenance and Attribution, to guard corporations from unintentionally including restrictively licensed code into their codebase when utilizing generative AI to write down code.
The brand new characteristic checks AI-generated code towards public GitHub repositories to seek out matches, and flags the license kind of the unique code from GitHub.
In line with Tabnine, this new characteristic will assist software program growth groups perceive if the code generated by AI fashions meets their requirements and necessities.
“State-of-the-art LLMs like Claude 3.5 Sonnet and GPT-4o have significantly improved the efficiency of generative AI functions, together with AI code assistants. Nevertheless, these LLMs are educated on huge quantities of knowledge collected from all corners of the web, together with code that will have restrictions on how it may be used, introducing the chance of IP infringement. Because the copyright legislation for using AI-generated content material continues to be unsettled, engineering groups at enterprises wish to strike a stability: leveraging the efficiency good points that come from these highly effective fashions whereas minimizing the chance of copyleft-licensed code getting of their codebase,” Tabnine wrote in a weblog put up.
Tabnine had already supplied a license-compliant mannequin educated solely on permissively licensed code, however this new characteristic will allow customers to leverage a wide range of different fashions too, similar to Anthropic’s Claude, OpenAI’s GPT-4o, and Cohere’s Command R+.
The Code Provenance and Attribution functionality additionally helps different growth actions on prime of code technology inside Tabnine, together with fixing code, producing take a look at instances, and implementing Jira points.
The corporate can be engaged on increasing this functionality to allow customers to specify repos to verify towards, similar to a competitor’s code. It should additionally add a censorship functionality that removes matching code earlier than the developer sees it.
Code Provenance and Attribution is at present out there as a non-public preview for all Tabnine Enterprise clients. Tabnine will even host a webinar on January 9 at 11 AM ET / 8 AM PT to dive into the potential additional.