GitHub Copilot for business ‘expected to boost enterprise adoption’

GitHub Universe sign outside the entrance steps of the Yerba Buena Arts Center

Open source software development platform GitHub has announced Copilot for business, which is expected to see more companies and large enterprises adopt the technology.

The new initiative introduces purchasing and provisioning methods that aim to make buying licenses easier for development teams, with a view to increasing the one million users Copilot has today.

Copilot for business will allow companies to bulk buy licenses and assign policies over which users within a company can use them.

Previously only available for purchase by individuals, it replaces the issue where organisations would have to assign a stipend to developers to buy Copilot for themselves and then expense it back to the company.

“We definitely had a lot of companies doing that,” said Jeremy Epling, VP product at GitHub to IT Pro. Epling also said he had encountered companies whose entire development teams bought Copilot individually, making the procurement process difficult.

Other large enterprises also told GitHub they'd been waiting for this announcement before adopting Copilot across the business so wider uptake is expected, according to Epling.

Copilot for business will not become cheaper for businesses to buy, Epling said, but the purchasing and managing of developers using it will become more straightforward.

Copilot has seen a wide uptake by developers due to the way in which it simplifies programming. Copilot suggests lines or even entire blocks of code, such as functions, so developers can focus more on tackling more complex code issues.

GitHub claimed 30% of all code written in repositories hosted on the platform is being written using Copilot and the number jumps to 40% for specific languages such as Python.

The announcement comes as GitHub was hit with a class-action lawsuit on 3 November regarding Copilot’s alleged “software piracy on an unprecedented scale”.

The thrust of the accusations relates to how Copilot’s artificial intelligence (AI), developed by Codex, is trained on public codebases found on GitHub.

The use of code generated by Copilot, which suggests code using the learnings made from scanning public codebases, could allegedly violate the open source software licenses that protect the public repositories from which Copilot is trained.

GitHub’s CEO Thomas Dohmke said in a press conference at GitHub Universe that GitHub added a feature to Copilot that enables developers to exclude code taken from public repositories in Copilot’s automatically generated suggestions.

The company has also added a feature to its code editor that allows developers to see from which repository Copilot-suggested code has originated.

“There's a lot of challenges and things to overcome,” Epling told IT Pro. “In this case, even potentially legal things so, I think for us, we're very bullish and excited about the future of AI, and how can be transformative for developers, and are excited to help make that known to everyone and make it a great experience.”

Asked about the company’s advice to businesses wary of adopting Copilot in relation to potential copyright issues, Epling said GitHub is working with customers to communicate what it’s doing with AI and Copilot.

Epling also said customers are split between either expressing a willingness to adopt Copilot immediately and those who will “wait and see”.

What is GitHub Copilot?

Announced in June 2021, GitHub Copilot is a tool that uses AI to suggest code to developers, allowing them to automatically generate functions or easily understand how to programme using an unfamiliar API.

GitHub’s narrative revolves around taking long, repetitive tasks away from developers so they can focus more on the core logic behind applications rather than expending brainpower on writing boilerplate code.

The tool has seen a wide-scale uptake from the developer community and GitHub’s research has suggested that it can significantly reduce the time developers need to build applications.

Research from the company pitched two dev elopement teams against one another, instructing them to build a web server using JavaScript. Developers who used Copilot showed a higher completion rate than non-Copilot users and did so, on average, 90 minutes faster.

GitHub also said that it improves developer happiness by leaving them to solve more complex problems and stay in a development flow for longer.

Instead of searching through a new API’s documentation before writing code, Copilot offers suggestions on how to use it, minimising the time spent outside of the developer environment, for example.

The AI that underpins Copilot is trained using code from publicly available repositories on the GitHub platform and the legal challenge brought against the company last week highlighted a potential violation of open source software licenses that protect code in such repositories.

Introducing ‘Hey, GitHub’

GitHub also announced the launch of its voice-powered augmentation for Copilot aimed at increasing the accessibility of software development to those who cannot use a keyboard.

Hey GitHub is a new feature that acts similarly to how voice assistants like Siri and Alexa work. Programmers are now able to input commands using their voice and Copilot will generate the desired code.

Currently in beta via a waitlist and accessible only in VSCode for now, the tool has already benefitted a team competing in a Hackathon event, GitHub said, as one of their team members was unable to type due to repetitive strain injuries.

It’s still early in development but the company thinks it will benefit a wider pool of developers beyond the disabled, and may benefit those with temporary injuries, such as a broken wrist, to remain in work if they want to.

Connor Jones
Contributor

Connor Jones has been at the forefront of global cyber security news coverage for the past few years, breaking developments on major stories such as LockBit’s ransomware attack on Royal Mail International, and many others. He has also made sporadic appearances on the ITPro Podcast discussing topics from home desk setups all the way to hacking systems using prosthetic limbs. He has a master’s degree in Magazine Journalism from the University of Sheffield, and has previously written for the likes of Red Bull Esports and UNILAD tech during his career that started in 2015.