Readers of this space are by now intimately familiar with the Device for the Autonomous Bootstrapping of Unified Sentience (DABUS), the machine created by computer scientist and prolific inventor Stephen Thaler that purports to have invented several items on its own.

According to Thaler, DABUS independently developed an improved liquid container with enhanced surface area for insulation and gripping ability, as well as an improved emergency beacon. Thaler applied for patents around the world in DABUS’s name, claiming in the “Inventor” section of the application that “the invention was autonomously generated by an artificial intelligence.”

Thus far, however, Thaler’s quest to achieve inventor recognition for DABUS has mostly fallen flat. Of the nearly 20 jurisdictions around the world where the patent applications were filed, only South Africa thus far has obliged; several other patent offices are still weighing the application, but the vast majority have held that only humans can invent.

As we have previously explored, the debate over machine-generated inventions largely tracks the distinction I have drawn between autonomists and “automatoners.” The former believe that robots have achieved, or will soon achieve, an important measure of independence from their human programmers and can genuinely be considered autonomous. In contrast, automatoners generally regard AI as a mere extension, however sophisticated and powerful, of human invention, incapable of independent thought or creation.

Two recent developments in the DABUS case have bolstered the automatoner viewpoint, but with a twist.

First, in May, a Tokyo district court affirmed the Japanese patent office’s rejection of Thaler’s application on behalf of DABUS. Judge Motoyuki Nakashima held that Japan’s Intellectual Property Act presumed human inventors: “The term ‘intellectual property’ as used in this Act means inventions, devices, new varieties of plants, designs, works and other property that is produced through creative activities by human beings.” (Emphasis added.)

Nakashima wrote that “many countries are cautious about interpreting the phrase ‘inventor’ in the patent laws to immediately include AI.” However, the judge left the door open to AI-generated invention as the technology matures. According to the summary by the Tokyo-based Taiyo, Nakajima & Kato law firm, one of the leading patent practitioners in the country, the court held that, “in the future, in consideration of the differences between the autonomous creative capacity of AI and the creative capacity of a natural person, it is not inconceivable that a system for the protection of AI inventions might be designed” and that “it is crucial to thoroughly examine the issues related to AI inventions via legislative debate in Japan and to determine specific conclusions without further delay.”

Then, one month later, the German Federal Court of Justice (BGH) similarly ruled that only humans can invent under German law. “Only a natural person can be an inventor within the meaning of Section 37 (1) of the Patent Act,” the BGH concluded. “A mechanical system consisting of hardware or software cannot be named as an inventor, even if it has artificial intelligence functions.”

Like the Japanese ruling, the German one saw human action behind every advance made by AI, insisting that

the fact that an artificial intelligence system has made a significant contribution to the discovery of a technical teaching does not contradict the assumption that there is at least one natural person who is to be regarded as an inventor due to the contribution he or she has made.

Unlike the Japanese court, however, the BGH did not care to speculate on whether future AI developments might warrant inventor status, or future legislative action might confer it.

Thus, the automatoner perspective has notched two more recent victories, even as the game plays out.

The post AI-Generated Inventions Suffer Two More Setbacks, Bolstering the “Automatoner” Viewpoint appeared first on American Enterprise Institute – AEI.