DABUS. Can Artificial Intelligence file for a patent?
Current legislature on patents and intellectual property design and development are limited in the area of whether machine is allowed to file patents. Various IP regulatory bodies are now working on understanding if Artificial Intelligence is allowed to own patents, trademarks and the intellectual property it develops through artificial means mimicking human behavior and thought. In recent news US regulatory law has blocked the ability for machine to invent and to claim ownership of an invention it is responsible for. U.S. District Judge Leonie Brinkema in Alexandria, Virginia has done just that by indicating that an invention cannot be owned by machine rather must be filed by a person and ownership traced back to a human.
The topic proves of importance in that as AI improves or advances, it’s ability to replicate or enhance human decisions may change various aspect of life, one being the development of new tools, objects, methods, designs and inventions. The USPTO has assessed the situation by understanding the circumstances, but will reject applications not submitted by humans—an individual must be the applicant stating AI has not reached the point where is ready to claim ownership on IP.
In a globally shared world where all types of skill sets work individually or collectively to produce or improve upon new things, the topic of non human inventors is very important in that when AI becomes capable of understanding the confounds of logic, it’s environment and how to modify the elements around it to produce innovation, kick starters, engineering and scientists are arguing that filings and ownership should be valid by machine. In a case called DABUS (Device for the Autonomous Bootstrapping of United Sentience)—an intelligent device with attributable strength in creativity—the machine was able to produce a beverage container. The application was rejected in the United States under the argument that AI is not at a point where it can be claimed by machine by artificial means.
Similar cases may arise soon as people work on AI technology day to day and improve rule sets and algorithms mimicking human behavior and decisions. Given the topic still needs further analysis and understanding, we do think that the USPTO’s inclination that “AI is not at a point to invent” is correct. Until we are certain for a fact that AI has the independent and cognitive ability to understand it’s inventions, these filings should be rejected unless an advanced design were to come out that collective bodies unanimously understand and approved on. Some attributes to validate this could be:
1. Machine learning adjusts algorithmic schemas based on use case (for what purpose the invention is being created) independent of the machine’s designer’s assistance. And if an assistance were to occur, it would know why such change is necessary.
2. Machine cognitively understands the use case for the intended invention.
3. Machine cognitively understands the historical evolution of the intended invention as predecessors may exist in different shapes and forms.
4. Machine simply understands why it’s inventing.
Given the above, a machine may still produce profoundly qualitative products, and could possibly even combine, align, realign and mix and match attributes far better than human ability, but it’s actions are a byproduct of human coding and editing—the machine’s designer has implemented the prerequisites to carry out specific functions. That itself traces novelty back to individual. In the example of DABUS, machine put together elements that collectively produced a “beverage container,” but raises questions on does that machine understand thirst, why the container is shaped the way it is to accommodate the hands, as well as, the the shape of the lip to accommodate the human mouth, including external and internal depths of wall thickness to safely house the container’s content.
It’s safe to say until AI understands it’s purpose from a cognitive and behavioral standing, IP filings should remain human. However, nothing prevents an individual to use his or her own AI design to help themselves improve calculations and provide better feedback on what’s being invented to file—kind of like a calculator.