As a postgraduate student with an interest in empirical legal research and technology regulation, attending Professor Heald’s public lecture [Blog link] on recent developments in empirical research on intellectual property law was inspiring. He didn’t just share research – he showed us how data is being used to help policymakers make better decisions about IP laws, and how we can critically draw insights from different empirical studies and learn what to look for when conducting our own empirical studies. The full recording of the public lecture can be found here:
One of the things that struck me was how straightforward Prof Heald was about the problems with so-called “empirical” research. It seems that although some databases claim to contain empirical legal studies, many of the papers in them don’t actually use any data, which is very surprising from my point of view. In this context, he also explained why we need more data-based research in IP. He argued that we already have a lot of economic theories, and now it’s time to test whether these theories actually match what happens in the real world.
Among the papers he presented, one example that caught my eye was a study where researchers trained an AI model (BERT) to predict whether the US Patent and Trademark Office (USPTO) would find a trademark to be inherently distinctive. The model predicted the office’s decision correctly most of the time, but sometimes it disagreed with the USPTO. This makes me think of a bigger question: since AI is good at finding patterns, does this mean that AI decisions are sometimes more accurate than human decisions, if they are properly trained and if people really want to strive for accuracy? In that case, how should we deal with the relationship between AI and enforcement? (Something for technology law scholars like me to work on).
Another project that I found fascinating was Prof Heald’s toaster study. He and his team scraped images of toasters from Amazon in the US, France, Germany and the UK. Using AI, they compared the variety of toaster designs in each country. The idea was to see if stronger design protection laws in the EU led to more diverse products. They found more variety in France and less in the US, which seemed to support the idea. But then came the big problem: what if this difference was caused by consumer taste, not the law? He admitted that they couldn’t prove that the law was the reason. As a student, it was really helpful to see a professor talking openly about the difficult question of ‘causation’ for most empirical studies. It showed me that not every research question has a clear answer, and that’s still valid and credible empirical research.
He also shared a clever study on disclaimers in trademark law. In this study, researchers tested how people reacted to fake ads selling vintage Notre Dame T-shirts. Some ads had no disclaimer, some had a small disclaimer, and some asked people to type in a clear disclaimer. The results were clear: the more obvious the disclaimer, the more action the consumer had to take during the disclaimer process, the less confused people were. This was very interesting from my perspective because the conventional wisdom is that disclaimers don’t work, but the experiment showed that they can if the process is designed with enough psychological tricks. This is particularly inspiring as it got me thinking about how law enforcement can benefit from behavioural design that helps to avoid litigation, which echoes and contrasts sharply with the legal concerns discussed in the previous public lecture by Amelia Fletcher on ‘Implications of behavioural economics for the pro-competitive regulation of digital platforms’ [Student reflection link] [YT link] who showed us how behavioural design in online choice architectures can, in the wrong hands, cause problems for effective regulation.
Prof Heald also talked about copyright and how what isn’t enforced can be just as important as what is. He gave the example of Jacques Brel songs on YouTube being uploaded without permission but not taken down. This could be because keeping the music online helps keep the artist in the public eye. Such an idea also ties in with a recent paper by Kristelia García, who also gave a public talk at CREATe on ‘Refusals to License and other Vertical Mischief in Copyright’ [Blog link] [YT link]. These findings resonate with my real-life experience as a fan of various niche pop culture, where I see idol group companies deliberately avoiding strict copyright enforcement on certain works as a way of promoting business.
Overall, Prof Heald’s talk was more than just a look at recent research. It showed how empirical methods can help us better understand the real-world effects of IP laws and more.