Higher AI content in M&E raises legal risks

A UK court ruled in favor of Stability AI against Getty Images in a copyright case, signaling challenges for copyright owners as AI litigation rises.

A UK high court recently ruled in favour of the London-based artificial intelligence firm Stability AI in a copyright infringement case in which international photo agency Getty Images had claimed the former had copied millions of its images. Getty’s pictures were used to train the Stability AI model and the ruling is seen as a setback for copyright owners. Last week also saw online marketplace Amazon filing a case against Perplexity AI in the US for its ‘agentic’ shopper that helps consumers choose the best and cheapest products online. It skirts Amazon’s standard shopping interface and skips display ads and sponsored product lists causing ad revenue loss. This, experts said, could have implications for its entertainment site Prime Video, too.

Play Button Media Player Podcast Graphic

Though there are no similarly high-profile AI related cases in India in the last week, AI content related litigations in the media and entertainment sector are set to rise. “As more AI-generated content is out, the number of lawsuits will increase. Cases by music labels are already pending in the Indian courts claiming their copyright music content was used for training AI,” said Akshaya Suresh, partner at national law firm JSA.

An expert in technology laws, AI, privacy and data protection, Suresh said that the row with Perplexity could affect Prime Video too. “Amazon has said that Perplexity’s AI agent is messing with its algorithms and bypassing its technical safeguards in e-commerce. Nothing stops it from altering recommendations for Prime Video audiences affecting viewership and eventually the economics of the business,” she said.

Rajiv Dingra, founder and CEO of ReBid, an agentic AI agency which has recently launched an AI Creative Studio, agreed that the Perplexity dispute has implications for streaming and advertising ecosystems. “Streaming platforms monetise via subscriptions, ads and content recommendations. If AI agents bypass or manipulate the user interface by selecting shows, bypassing ads, and recommending outside the platform logic, that may threaten monetisation models,” Dingra said.

Back home, where AI regulation is still at a guidelines stage, the government’s proposal to clearly label all AI-generated content or synthetic media (text, image, audio, video) with visible markers, has caused a commotion.

“The creative community (advertising, film, design) flags that such broad labelling may hamper creative freedom, impose compliance burden, dilute aesthetics, raise costs, and create ambiguity (such as, what counts as “AI-generated”?),” said Dingra. If every piece of AI-generated content gets a visible “AI-generated” label, brands may feel it devalues the content or signals second-class creatives, which could impact viewer perceptions, he added.

There may be implementation challenges too. “Platforms must build detection or verification systems; creators must embed metadata and labels. In a price-sensitive market, this may increase production cost or reduce flexibility. It may also stifle smaller creators, startups and experimental art/entertainment formats,” Dingra said.

However, mandatory labelling of synthetic content is a way for India to address the problem of deep fakes, said Akshaya Suresh. She said even globally there is consensus on the issue. Dingra, however, believes that its implementation must balance clarity with creativity and not inadvertently penalise responsible AI use.

Also, if AI can cheaply assist or produce “creative” work, then the business model of agencies, creators or licensed-assets may get squeezed, he said. Who owns the output, who gets paid, how is IP valued are some questions that will come up, he noted.

Issues around legal and ethical compliance will rise as most AI models get trained on large datasets of images, video, music etc without clear consent or licensing. UK’s recent ruling on Getty Images versus Stability AI sends a signal that rights-holders are ready to litigate, Dingra said. However, Suresh pointed out the difficulties in proving copyright infringement. “Getty was not able to substantiate it through evidence as AI has its own way of learning… it does not copy the same content and store it,” she said.

For India, Suresh said, overregulating can stifle innovation and create barriers for AI companies to do business. The government’s approach in tackling the most serious harms first --- like deep fakes – is measured, she said, adding that regulators like RBI and SEBI are creating contextual AI frameworks for their sector. “This crawl-walk-run approach is good for a developing economy like India,” she said.