Tabnine, a leading AI coding assistant, has introduced a groundbreaking Provenance and Attribution capability to support the use of large language models in software development while minimizing intellectual property (IP) liability. This innovative feature enables enterprise developers to leverage AI-generated code while ensuring compliance with licensing restrictions.

The new capability, currently in private preview, allows Tabnine to check code generated using AI chat or AI agents against publicly visible code on GitHub, flagging any matches and referencing the source repository and its license type. This detailed information enables engineering teams to review AI-generated code and determine whether the license meets specific requirements and standards.

The need for such a capability arises from the tradeoff between using larger, more performant language models trained on broader datasets and the risk of IP and copyright violations. According to Tabnine President Peter Guagenti, models trained on larger pools of data outside of permissively licensed open-source code can provide superior performance, but enterprises using them risk running afoul of IP and copyright laws.

The Code Provenance and Attribution capability addresses this tradeoff, increasing productivity while maintaining compliance. This proactive stance is particularly important in the current legal landscape, where copyright law for using AI-generated content remains unsettled. By reducing the risk of IP infringement, Tabnine's feature supports enterprises using models like Anthropic's Claude, OpenAI's GPT-4, and Cohere's Command R+ for software development.

The capability supports a range of software development activities, including code generation, code fixing, generating test cases, and implementing Jira issues. Looking ahead, Tabnine plans to expand the feature to allow users to identify specific repositories, such as those maintained by competitors, for generated code checks. Additionally, the company aims to introduce a censorship capability, enabling administrators to remove matching code before it is displayed to developers.

This development marks a significant step forward in the responsible use of AI-generated code in enterprise software development. As the industry continues to grapple with the implications of AI-driven innovation, Tabnine's Provenance and Attribution capability sets a new standard for minimizing IP liability and ensuring compliance in the age of AI-assisted development.

With this capability, Tabnine reaffirms its commitment to supporting development teams and their legal and compliance counterparts in leveraging the power of AI while maintaining the highest standards of intellectual property protection.