GitHub Copilot, a service provided by GitHub that intelligently suggests lines of code, was previously available for individual users and educators. After months, it’s now available in a plan for enterprises.
The new version of Copilot, called Copilot for Business, comes with all the standard Copilot features for individual users. In addition to those, it includes corporate licensing and organization-wide policy controls.
The plan will cost $19 per user monthly. Promising industry-leading privacy, GitHub adds that the offer:
Copilot, which has had over 400,000 subscribers since August, can create a solution or a programming approach to help developers accomplish what they describe to the AI, thanks to its knowledge base and context.
The developers can receive suggestions from GitHub Copilot by writing the code they want to use or by writing a natural language comment describing what they want the code to do.
Even though they used to be a Sci-Fi concept, Artificial Intelligence, and machine learning are all around us today.
The AI model that powers Copilot is called Codex, developed by OpenAI and trained on billions of lines of public code. Copilot is a downloadable extension for development environments, including JetBrains, Microsoft Visual Studio, and Neovim.
The senior director of product management at GitHub, Shuyin Zhao, says that they sped up the rollout of their new offer after hearing:
However, it’s uncertain if businesses will rush to purchase their product, as Copilot is amid unresolved legal issues regarding copyright infringement.
GitHub’s Legal Issues Over Copyright
Some advocacy groups have taken issue with the AI because at least a part of the code Codex was trained in is copyrighted or under a restrictive license. Users noticed Copilot generating questionable code months ago.
Tim Davis, a computer science professor at Texas A&M University, spotted the tool producing,
When public code was blocked using a filter introduced to Copilot by GitHub in June. According to a document from GitHub, it:
The feature seems to be imperfect, according to the discovery made by Tim Davis. Earlier, users could obtain example code from books like “Mastering JavaScript,” code snippets from personal databases, and even generate code from Quake, using Copilot.
GitHub warns that about 1% of the time, a suggestion may contain code snippets longer than ~150 characters that match the training data.
GitHub says that the tool can produce code with “undesirable patterns,” which may put its users at risk. In its document, GitHub warns its users of possible lawsuits if overlooked copyrighted suggestions make it to the final version of their product.
Programmer and lawyer Matthew Butterick partnered with the class-action firm Joseph Saveri Law Firm to accuse GitHub of violating copyright law by allowing Copilot to recreate sections of copyrighted code without providing credit.
The class-action lawsuit filed in November targets GitHub, OpenAI, and Microsoft.
With plans to introduce additional features in 2023, GitHub aims to help developers make informed decisions regarding Copilot’s suggestions, including the ability to identify the repositories where used public code originates.
GitHub claims it won’t retain code snippets or share code, regardless of whether the data comes from private repositories, public repositories, non-GitHub repositories, or local files.
Despite GitHub’s determination, it’s unclear if companies will be willing to take the risks that using Copilot brings.
Read More Software News:
AI Developer Jasper Valuation Jumps to $1.5 Billion in Latest Funding Round
Best Project Management Software Comparison — Top 12 Tools for 2023