How Courts Are Ruling On AI's Rights And Wrongs

Artificial Intelligence products are currently working in a regulatory vacuum in India. The government rolled out a quick advisory on March 1, telling platforms to label AI models that haven't been fully tested as "under testing" or "unreliable". It also said these products would need government approval before being released to the public.

This prompted a backlash from stakeholders, calling it a return of "license raj".

Eventually, though, the government pulled back on the advisory.

Worldwide, too, regulators are struggling to come up with a policy framework that can be dynamic in governing the evolving AI models and yet not hurt innovation.

The U.S., Australia, China and Singapore lack a federal law specifically addressing AI and its implications. Europe has an AI Act, but it focuses on the technology's output and the risks it poses. Japan has opted for a softer approach, emphasising ethics and safety in AI applications.

Lack of clear answers on AI's consequences in the commercial world has prompted affected parties to knock at the doors of regulators and courts. 

Just last year, in two separate cases, the U.S. Copyright Office and the U.K. came to the same conclusion on the question of ownership of AI outputs.

In the case of graphic novel "Zarya Of The Dawn", the Office ruled that copyright registration will apply only to material created by the author and not to graphics generated in it by the AI.

The U.K. Supreme Court issued a similar ruling in the case of AI researcher Stephen Thaler, who filed two patent applications. One, to secure a patent for his AI-powered machine, DABUS, as an inventor. And two, for any invention made by DABUS. The court denied both saying an inventor has to be a natural person, and that there's no basis for granting him a patent for inventions made using DABUS.

Copyright battles are playing out in courts not just for AI outputs, but inputs as well.

The New York Times sued OpenAI and Microsoft, saying they used millions of its articles to teach AI chatbots. Author Julian Sancton and several others brought a class action suit against OpenAI and Microsoft on grounds that ChatGPT was trained on the works of their copyrighted material without prior permission.

Coders, too, are an aggrieved lot.

In November 2022, a group of programmers sued GitHub, Microsoft, and OpenAI in a California court. They said these companies violated their copyrights by using their code to train AI coding tools, named Codex and Copilot. The programmers claimed GitHub used their code without permission, breaking open-source licences that require giving credit and getting permission for commercial use. The court is yet to give a ruling in this case.

Another pending question before courts is AI's implications for privacy.

A Californian court will decide on a class action suit alleging that in developing and marketing ChatGPT and its later versions, Microsoft and OpenAI stole private information, including personally identifiable information, from hundreds of millions of internet users—including children of all ages—without their informed consent or knowledge. The lawsuit says the companies gather various kinds of data, like user locations, Snapchat images, financial details from Stripe, music tastes from Spotify, chats from Slack and Microsoft Teams, and health information from patient portals like MyChart.

Indian courts are yet to see any of these legal conundrums. Though the Ministry of Commerce is confident that the Copyright Act, 1957, is well-equipped to protect AI-generated works and there is no need to create separate category of rights.

The exclusive economic rights of a copyright owner obligates the user of Generative AI to obtain permission to use their works for commercial purposes if such use is not covered under the fair dealing exceptions, Union Minister of State for Commerce and Industry Som Parkash had submitted in a written reply to the Rajya Sabha last month.