It didn’t seem that long ago that people were talking about Artificial intelligence or AI, as something futuristic, or another generation away. Not anymore. AI is rapidly infiltrating many aspects of our lives with automated assistants, smart devices, driverless cars, toys, and more.
While there are seemingly limitless benefits of AI, there is high-level concern about AI’s impact on people and jobs. Research group Gartner predicted that by as early as next year more than 3 million workers globally will be supervised by a “robo-boss.” There is no denial that this technology is creating such unprecedented disruption that governments, companies and societies around the world are racing to manage and control AI before AI controls us.
Even the IP industry isn’t immune. And it’s not just the filing of patents for AI technologies that’s of concern. AI raises intriguing questions such as who the inventor is where the machine invented or contributed to inventing the technology. From an IP industry perspective, questions are being raised about stated initiatives from IP firms and patent offices such as the EPO to conduct patent searching, examine patents and even write patent specifications using AI.
Identifying the inventor – human or robot?
A relatively easy initial target for AI may be patent searching since this is a time-consuming task. A patent search usually has some defined scope and often search terms and strategy evolve as a search progresses based on results found and learnt from. AI could ease the work involved here or at least supplement the efforts of real people. But what would happen if the AI system itself invented something, for example in the process of searching or in the development of a new product?
According to Ryan Abbott Professor of law at the University of Surrey School of Law, AI systems have been making patentable inventions for over 20 years, but patent law hasn’t kept up.
He said, “Soon computers will be routinely inventing, and it may only be a matter of time until computers are responsible for most innovation. To optimise innovation – and the positive impact this will have on our economies – it is critical that we extend the laws around inventorship to include computers.”
Machines inventing things contrasts with a key reason for patents being to recognise inventors (people) for their inventions in a public forum (a patent office).
Meanwhile, the number of patents issued around the word for AI enabled technologies has been skyrocketing in recent years (see IPWatchdog, VOX, iProgrammer). In the US alone, market intelligence firm Tractica released a market taxonomy of 27 industry sectors featuring potential AI use-cases and forecasted that global AI revenue would grow from $643.7 million in 2016 up to $38.8 billion in 2025.
But the question as to who would be liable for malpractice if the product or process was invented by an AI machine is another issue yet to be defined. Is the patent owned by the machine, the owner of the machine, or the person who programmed it? This is not a far-fetched question as legally, inventors have the right to file a patent unless other contractual obligations exist. Should an AI robot be listed as an inventor? Is this legally right since patent law defines the inventor as a person?
In October, Saudi Arabia became the first country to grant citizenship to an AI robot developed by Hanson Robotics called Sophia. She also said that she wanted to have a child one day, presumably to create another AI personality like her – Sophia II. Will Sophia have the patentable right for her progeny, or will Hanson Robotics?
At present, the European Patent Office, which sets a high standard for patentability, states that it would depend on the ‘technical’ versus ‘non-technical’ features that contribute to the ‘inventive step’. A computer processing non-technical information in some non-obvious way may not be considered inventive unless there is a technical effect. However, these standards vary around the world, and the impact AI will have on IP rights is yet to be seen.
AI patent systems may be faster, but are you willing to bet your patent rights on them?
There’s also the tricky issue of ethics. AI technology is designed to mimic humans, so will we be able to rely on AI automated patent advice or patent examination to provide a fair and unbiased service?
In the short term, while the world grapples with the implications of AI, filing and protecting of patents will still need to rely heavily on experienced and connected IP experts. But these investments in new technologies will inevitably be offset by fee increases and service disruption, perhaps initially for specific tasks like patent searching.
The hope is, once AI systems become better trained and their accuracy and understanding of the many IP anomalies improve, having supporting AI technology to plough through time-consuming tasks like patent searches will save time and create other positives for the industry such as greater patent quality. It would stand to reason, therefore, that the cost of these specific tasks will lessen over time.
With so much change in the IP industry, will the personal approach of the experienced IP strategist eventually be replaced by an AI supported IP graduate in a mega-IP group with the funds to invest in this technology? Possibly, but probably not for some time, at least until Patent Attorneys and their robo-bosses can argue with confidence against an AI judge as to who or what invented the technical step! Perhaps obviousness rules should be amended to not reflect what is obvious to a person skilled in the art but rather what is obvious to a ‘person or machine’ skilled in the art?