The race is intensifying to build self-driving cars, and artificial intelligence research has become a hot commodity. Automakers like Toyota are cozying up to top schools, including with a recent $22 million grant to University of Michigan.
Researchers are rapidly developing vehicles that are either fully self-driving or use artificial intelligence to assist driving. Billions are being poured into AI, robotics, optics, and mapping research, and not just from Google, Apple and Tesla—major automakers, cities, even the federal government are deeply invested in how this all unfolds.
In one sense, it’s a race between Silicon Valley and the automakers, as the latter try to catch up with companies like Google, which employs or otherwise funds a small nation of software developers.
- Google’s Quiet, Multimillion Dollar Academic Research Program
- Google Grants Teaching Machines to See, Hear, and Learn
To close the gap, one of the many strategies car companies are employing is turning to academic research. As a result, we’re seeing a lot of funds headed to universities, and some tight relationships developing between research departments and car companies.
Toyota, in particular, has made shortening the distance between fundamental AI research and its line of cars a very high priority, establishing the Toyota Research Institute division with a $1 billion initial budget.
That project has set up shop next to three of the country’s best robotics and artificial intelligence schools, establishing joint research centers to work closely with Stanford, MIT, and the University of Michigan. Most recently, the company announced a $22 million grant over four years to UM for artificial intelligence research.
We’re seeing a number of these partnerships emerge lately, as corporations speed up deployment of new advances.
For example, Ford, Volkswagen and Toyota, along with four other tech companies, are each committing $300,000 to the University of California at Berkeley for an initiative called DeepDrive, funding up to 30 grad students a year for AI research. Companies get to give feedback on research proposals and gain quick access to implement new developments. Meanwhile, schools gain funding and a chance to validate their work in real-life situations.
The research is competitive, but also surprisingly collaborative between companies. For example, the University of Michigan, the state’s DOT, and 15 primary corporate sponsors funded the $10 million Mcity, a research hub for autonomous vehicles in Ann Arbor that includes a little fake city for test drives. Sponsors include Ford, GM, Honda, Nissan, and Toyota.
These companies can be a huge benefit to academic research departments, whose AI and robotics work is suddenly hotter than ever. But they can also pose a threat.
The most notorious example was Uber’s descent upon Carnegie Mellon University. The company set up shop next to the school and announced a new partnership, but shortly thereafter hired away something like 40 of the university’s researchers. Uber later made a $5.5 million grant to the school, largely perceived as mending fences, but the partnership has gone nowhere since. Uber has set up a smoother working relationship with the University of Arizona.
The connection between Toyota and its partner schools is less poach-y, but the boundary is thin. For example, two top professors going to work at the new Ann Arbor research center will keep their UM positions, as long as they devote 80 percent of their time to autonomous vehicle-related research.
We actually see a lot of these arrangements related to university fundraising, as schools pick up big funding from corporations that stand to benefit from their work in a certain area. We speculated about a year ago that the dash for autonomous cars would be a huge benefit to schools with expertise in these fields, but that it could also be a drain on their faculty. Turns out that some of the emerging arrangements are much more symbiotic, combining academic research with corporate R&D.
Of course, funding like this also raises the question of just how "charitable" a corporate donation is when the money is so clearly aimed at boosting the bottom line. Or to put the issue differently, why are taxpayers subsidizing R&D for the automative industry through the charitable exemption?
Before you get hot under the collar pondering that question, keep in mind that since R&D expenses can also be written off, this kind of philanthropy has no effect on what companies pay in taxes. Still.