Artificial Intelligence, simply defined, is the imitation of the human intellect or human behaviour. The emergence of Artificial Intelligence in our day to day life has become more pronounced. The question that arises is, whether our Patent laws are adequate to incorporate AI in its patentability criteria. Also, another unanswered question is how to address the liability of harm or damage caused by the AI to human society. The important question to be answered is who should be held responsible for such damage.
For a patent to be registered, the invention must be novel, non-obvious and it must-have utility. For a patent to be novel, the invention must not be known or used and must not be printed or published in a journal. The most difficult part of being proved is non-obviousness. For an invention to be non-obvious, it must have a distance from the prior art. There must be an inventive leap.
The Story of Self-Driving Vehicles
Self-Driving Vehicles would be commercially available by 2025, and the recent statistics show that the development in the innovation is fast and a transport revolution is expected to come. Looking at the patent applications in this field, it gives a unique insight into the race to innovate in smart, connected and automated vehicles. The crash of the Tesla’s driverless car and explosion of the SpaceX Falcon 9 rocket at its Cape Canaveral Launch pad are strong reminders that when artificial intelligence (AI) driven technology is advancing at a dizzying pace, then regulators cannot lag behind.
Shortcomings of the Present Law
All of us have experienced the benefits of AI in various forms like- email spam filtering, Facebook’s auto-recognition to tag your friends and family, web page translation, booking cabs, ordering food, paying bills, recharging phones, etc. Countries like, U.S. and others have recognized the problem and have started the formulation of new policy whereas laws in countries like India still lag behind where the ethical and regulatory implications of Artificial Intelligence has not yet been recognized.
Recently, the U.S. Copyright Office and the World Intellectual Property Organization (WIPO) held a joint event titled, “Copyright in the Age of Artificial Intelligence” (AI) at the Library of Congress in Washington, DC. The event explored how global copyright law and intellectual property law, as well as broader policy, may currently address AI technology, and included dialogue about changes that may be needed. Panellists also shared how AI is being utilized now and what future technology deployment and innovation may look like.
The main concern is that the current patent laws are inadequate to protect the AI system. For example, the current patent laws cannot be used to protect the compilation of data, Artificial Intelligence training sets or a programmer’s particular expression of source code. Further, looking at the incremental progress and machine learning process of the underlying algorithm, it is very difficult to describe the functions of AI as required for granting a patent. The learning of the machine in present days attempts to duplicate human intelligence by interacting with the world and receiving the corrective inputs. It is almost like teaching a child right from wrong by scolding him or praising him. AI machines also learn in similar binary corrections.
Dangers of AI
Another major concern regarding AI is the liability in case of harm or damage caused by it. In the Uber AV fatality case, during real-world testing of an autonomous car operated by Uber, Elaine Herzberg was struck and thereafter killed by the car. This was believed to be the first case recorded regarding pedestrian fatality involving an autonomous car with a human emergency driver behind the wheels. The accident took place at around 10 pm, when the pedestrian was crossing a road while walking outside the crosswalk. The Uber car and the diver did not notice the pedestrian crossing the road, which resulted in an accident. Right after this event, the Self-Driving cars were suspended, and an investigation was set up regarding the tragic accident. Real-world testing has since resumed.
In accidents involving autonomous vehicles, the vehicle is operated by computer and not human, therefore, the focus of reconstruction experts will be on the product itself and activity or inactivity of the people inside the car would be irrelevant while determining the cause of the accident. This will require the experts to examine the hardware and software malfunctions that are required to control the self-driving car. This will require expertise beyond the fields of mechanical engineering and biomechanics such as computer science, data analytics and programming.
For analyzing the issues involved in the following case, we have to identify the possibilities involved in this case regarding liability. Before we dive into the technical issues, we have to analyze whether, in this incident, people are fundamentally at fault. The individual crossed the road while walking a bike at that time without using a crosswalk. This situation might have created confusion for the internal system for identifying the damage caused. Though the car and the driver must have been able to see the pedestrian but is it possible that the liability is on the pedestrian in the present case? Perhaps, the human drivers might have been paying reasonable care to the road and most likely noticed the pedestrian, at least swerving or braking to avoid a last-minute collision.
Similarly, there is a possibility that the human driver is at fault. The car was not actually operating without any humans at all in the vehicle. In the U.S., recent laws require a human to be inside a moving car where they can control the steering of the car. In the present case, there was a driver inside the car when the autonomous mode was engaged. This is basically not fully autonomous mode with Level 5 operation. It still relies on humans for a backup. In this case, the human failed to provide that backup.
It is really reasonable to assume that a human can remain completely unaware of their surroundings and assuming that the machine would do all the work, and then be asked to step in at a critical moment to handle a life or death issue with very little notice? It is very difficult to see how instantly the human can be unaware to acutely aware in such a short span of time. Perhaps the entire assumption and expectation of the human backup driver are unreasonable.
The question of whether legal personhood can be conferred on an AI boils down to whether it can be made the subject of legal rights and duties. A possible middle ground may be granting AI a bundle of rights selected from those currently ascribed to legal persons. However, concrete steps in this regard are yet to be seen. Another issue that arises is attributing liability to an AI. The general rule has been that since an AI cannot qualify as a legal person, it cannot be held liable in its own capacity. The biggest roadblock to reconsider this rule is the conundrum as to how to penalize an AI for its wrongdoing, which has not been dealt with as of today.
References patent Eligibility requirements, (January 10th, 2020, 19:20 pm) https://smallbusiness.findlaw.com/intellectual-property/patent-eligibility-requirements.html.  Fagnant DJ &Kockelman K (2015) Preparing a nation for autonomous vehicles: opportunities, barriers and policy recommendations, 167–181.  Piao J & McDonald M (2008) Advanced driver assistance systems from autonomous to cooperative approach, 659–684.
This Article is written by Vanshita Jain, Final year law student at Institute of Law, Nirma University.
Libertatem.in is now on Telegram. Follow us for regular legal updates and judgements from the court. Follow us on Google News, Instagram, LinkedIn, Facebook & Twitter. You can also subscribe for our Weekly Email Updates. You can also contribute stories like this and help us spread awareness for a better society. Submit Your Post Now.