Featured Image Credit: John Greim/LightRocket via Getty Images
One of the biggest conversations surround the rapid progress made by AI companies is the data that they're training the models on, as many people have claimed that this process breaches copyright laws, seeking damages as a result.
While philosophers have proposed that artificial intelligence will never truly achieve sentience, it certainly can't think for itself right now and has to get all of its data and information from somewhere.
Some of what it feeds its users might be made up – leading to both hilarious answers in some cases and concerning results in others – but the vast majority of its knowledge and 'intelligence' comes from billions, if not trillions of training data sources.
Some AI companies have paired up with sites like Reddit in expensive partnerships to train their models on that data, yet other companies have claimed that their material has been used unlawfully and have looked for compensation off the back of this.
Advert
Several lawsuits have already been launched, including one recent legal challenge from a group of prominent YouTubers, but one of the biggest has now came from major news network CNN as it has launched its first ever lawsuit against an AI company.
As reported by Variety, the lawsuit targets AI startup Perplexity with CNN claiming that its copyrights and trademarks have been infringed after content was copied and distributed without the network's consent.
At the heart of the case is the claim that Perplexity scraped over 17,000 different CNN stories, photos, and videos to train its AI model, and seeks damages that could stretch into the billions of dollars if the AI company's $20 billion valuation is anything to go by.
The lawsuit's filing outlines that "Perplexity's latest valuation at $20 billion and success at raising funds of nearly $1.5 billion are indicative of the potentially massive illegal transfer of economic value from original content creators like CNN to Perplexity," adding that they "paid CNN nothing for using CNN's content to power its products" despite spending millions of dollars on cloud services, talent, and the outsourcing of its models.
Similar AI companies like Meta have worked out a content-licensing deal with CNN, showing the network's willingness to cooperate if a financial incentive is there, yet claims that Perplexity has 'exploited' the original content put out by the network for their own benefit.
It's safe to say that Perplexity doesn't seem too bothered about the legal threats made by CNN, however, as a spokesperson's four-word statement to Variety expresses their stance on the matter.
"You can't copyright facts," the spokesperson outlined, suggesting that the data used to train its models isn't necessarily 'original' or 'owned' by CNN but something that exists in its own independent state.
The company has expressed similar sentiments towards other lawsuits in the past, with a legal challenge from Dow Jones back in 2024 prompting the company to denigrate "an adversarial posture between media and tech that is – while depressingly familiar – fundamentally shortsighted, unnecessary and self-defeating."
