Artificial intelligence (AI) is proving pivotal in addressing the complex challenges faced by automotive manufacturers and mobility providers. Systems that can learn for themselves promise to rewrite the rulebook for insurance risk assessment, driver assistance and safety features, and traditional production and product development processes.
Mobility players are investing heavily in the space and—judging by the number of AI-related patent filings—interest is growing. But at the same time, the increased prevalence of AI within the mobility ecosystem could raise issues around intellectual property (IP) strategies. Some legal experts are warning that companies may need to become more strategic in how they protect their IP to maximise business value.
IP protection: a good fit
The World Intellectual Property Organisation characterises the period between 2012 and today as an “AI patent boom” driven by increased data and connectedness along with greater computer power. Of all AI-related patents filed, 42% were in the areas of transportation, telecoms, or life and medical sciences.
Transportation in particular has been powering ahead. The industry represented just 20% of AI applications in 2006. By 2016 it accounted for one-third with more than 8,700 filings. But when it comes to classification areas, the lines are blurred. “Sometimes it is hard to capture the true scope of what AI is, and further, what would be considered to fall within automotive,” cautions Pavan Agarwal, Partner at international law firm Foley & Lardner.
The trend, though, is clear: momentum is growing. Agarwal works with companies to determine what kind of IP best fits their need. That may involve obtaining a number of patents or protecting something as a copyright or as a trade secret; different aspects of AI are better suited for different forms of protection. Education is often a central feature of these talks. Some software engineers and computer scientists are less familiar with obtaining patents, and the sector has seen some debate over their value. Agarwal comes out squarely on the pro-patent side where it makes sense: “I believe patent protection is valuable to the software industry, and we try to help them understand that.”
When it comes to protection, the focus is on two key areas. The first is the core AI algorithm. “There may be some unique aspects to the nature in which they’re doing the AI itself,” Agarwal explains. The second area of protection is the application of AI to a specific area, such as automotive and autonomous vehicles.
But there are multiple routes down the IP path, and not everyone files patents themselves. Some seek to acquire patents from third parties. It often happens that a start-up company that hit the market a little ahead of its time is forced to wind down operations and decides to offer its patent rights. Companies may also obtain IP through joint development or supply agreements.
“It’s important to look at agreements that exist between collaborating companies,” Agarwal points out. The form of these collaborations and the companies involved in them have changed considerably over the years. “The automotive industry is quite different than it was a decade ago,” he tells Automotive World. “The traditional tier relationships have been disrupted by the entry of numerous technology players, big and small. It’s also been impacted by the change of the car from something primarily mechanical to something closer to a computer on wheels.”
Challenges and opportunities
One of the industry developments that could lead to IP issues down the road is the current cash crunch. The novel coronavirus (COVID-19) pandemic and the subsequent market downturn has strained liquidity and caused many players to re-examine funding allocation plans. “Companies may have seen their R&D cut and their budgets tightened,” Agarwal points out. “Sometimes it is not so easy to see how filing these patent applications will result in specific value, as opposed to spending the money on sales, which may be seen as directly bringing in immediate revenue. There is always a trade-off in terms of how many patents to file for. They are not inexpensive things to obtain, so that may be an area that proves a challenge.”
The good news for automotive players is that there is still plenty of white space left to capture in the patent landscape, despite the increased activity lately. This is particularly true for the autonomous vehicle space. “There are so many ways to come at the problems around autonomous driving, and even areas within that sector like pedestrian detection,” he says. “That allows forward-looking companies to hold a brainstorming session with the engineers and even the business guys to map out where they see the technology being developed. If you can get into a certain white space in that way, then you can file for some protection as you continue to develop that area.” In Agarwal’s experience, brainstorming sessions like this have been a valuable approach for companies to obtain IP rights in a relatively untapped area.
Looking ahead, that patent landscape should see considerable activity. Agarwal expects a continued acceleration in patent filings, particularly around the AI space within autonomous vehicle applications. “Autonomous driving relies heavily on data and unique ways of processing data very quickly. We will see tremendous growth in terms of companies developing and then seeking to protect their IP,” he predicts.
However, there’s a flip side to all this development momentum: increased disputes. “Within the autonomous vehicle space, both with AI but also with the connected car regime, we will see more disputes between companies, suppliers, patent owners and perhaps even non-practising entities that want to obtain patents and seek to assert them,” he cautions. “As the area and the revenue associated with it grows, that tends to emerge as a by-product.”