EDITOR’S LETTER

AI no longer love you?

“It seems unlikely that it will succeed at any of the world’s major patent offices.”

Over the summer and into autumn, we heard a lot about the artificial intelligence (AI) inventor project known as “DABUS”.

In case you hadn’t heard, physicist Stephen Thaler, backed by Ryan Abbot, a professor of law and health sciences at the UK’s University of Surrey, is gamely ploughing on with a global project to get a patent for his “creative AI paradigm”.

In late July, the team managed to convince an Australian judge to overturn IP Australia’s rejection of Thaler’s application (an appeal is pending). This was shortly after South Africa “awarded” the machine a patent, which raised a few eyebrows given it was more of a bureaucratic victory than a substantial one.

However, in September, a US district court followed the opinion of the patent offices in the US, UK and Europe by saying that an AI cannot invent because, among other reasons, it doesn’t qualify as an “individual” under the America Invents Act.

Judge Brinkema also said: “As technology evolves, there may come a time when AI reaches a level of sophistication such that it might satisfy accepted meaning of inventorship. But that time has not arrived,” and the decision would then fall to Congress.

A September decision at the English Court of Appeal went the same way, albeit with an interesting dissent by Justice Birss, which requires further analysis and will no doubt give the project something of a boost amid the bad news.

Meanwhile, an excoriating blog on IPKat by AstraZeneca patent attorney Rose Hughes titled “An AI inventor or the emperor's new clothes?” asked whether we could all “move on” from DABUS, prompting lively agreement in the comments.

Given the project’s record to date it seems unlikely that it will succeed at any of the world’s major patent offices. And public interest in its journey is waning with every new rejection.

Whether it was ever likely to succeed was not DABUS’ entire purpose, of course. It raised questions that patent law continues to try to answer. But its entry into the history books seems to be arriving a little sooner than its creators expected.

Tom Phillips is the editor of WIPR

Image: shutterstock.com / Willyam Bradbury, chanchai howharn