World Intellectual Property Review quotes Dewdney William Drew on the Getty Images vs Stability AI decision

According to World Intellectual Property Review, the long-awaited Getty Images v Stability AI judgment landed with more of an anti-climactic thud than a bang last week.

After three years of litigation and weeks of hearings, many hoped the 205-page ruling would go some way towards answering the pressing question: does using protected content to train an artificial intelligence (AI) model amount to infringement?

It didn’t, leaving many to speculate that the time is ripe for the government to step in and provide guidance.

But the decision does offer a silver lining for rights owners, providing strategic pointers for those mulling legal action over the use of their trademarks in generative AI outputs.

Dewdney Drew, Head of Brand Protection, shares his thoughts with World Intellectual Property Review: 

Getty’s narrow trade mark infringement win over Stability AI is a disruptive proof-of-concept that may complicate AI model development.

"The case is somewhat unusual because the watermarking of digital images with trade marks is not universal.  It may be said that Getty themselves pioneered and popularised this strategy and so perhaps they were always in an almost unique position to adopt this line of attack.  Much of the attention in the IP versus AI debate has focussed on copyright, rather than trade marks.  It would however seem that the inclusion of a trade mark ‘spanner’ into the AI ‘works’ can cause it to jam in infringing ways.

"Given that the infringement findings are historic and extremely limited in scope, it remains to be seen what damages, if any, can be recovered but the positive finding establishes a new risk factor that AI developers will now need to take into account in training their models.

Read the full piece in World Intellectual Property Review here.