1. While automated vehicles should significantly reduce crashes and deaths behind the wheel, lower emissions, and bring down the cost of auto insurance rates, the technology raises a significant number of questions like:
• How quickly will consumers accept automated vehicles as part of their lifestyle?
• Who is liable in the event of an accident?
2. Allocation of liability among the stakeholders in the automated vehicle space will focus on design defects, software licensing disputes and the safety of the algorithms generated by artificial intelligence.
3. There is a shift already occurring in the insurance space by which rates are now being set based on how one drives. Insurers may decide to provide discounts to drivers who drive with their cars in an autonomous mode.
4. As an ethical question, we must ask whether artificial intelligence’s algorithms will be allowed to coldly calculate a cost-benefit analysis based on liability when choosing between two poor outcomes.
5. Information is power. As original equipment manufacturers (OEMs) advance toward highly automated vehicles (HAVs), they will gain access to a massive data supply, giving OEMs and their tech partners a significant advantage in monetizing the data, while challenging the transparency of privacy policies.
6. As certain OEMs or technology companies begin to roll out fleets of HAVs that people will hail, but will not own, the insurance model is going to have to adapt to deal with all potential models (hailing, buying, leasing). Liability for accidents is expected to shift from drivers and car owners to OEMs, suppliers and the new tech companies.
7. Connected and automated cars bring to the forefront a number of cyber risks. The various components that make up an automated car come from different organizations, making it harder to thoroughly test “the whole” for vulnerabilities pre-launch. In turn, certain issues may not be addressed until the vehicle is live and on the road.
8. Patent filings show an extraordinarily diverse marketplace; no one OEM or software supplier dominates the automated car industry as of yet.
9. Two methods of building artificial intelligence: Programming a computer, which has been the dominant method over the last half century, and machine learning, where the machine learns from its own experiences.
10. Unlike other revolutionary technologies in the past where a single company could produce the product from end-to-end, like the smartphone, no one company at this time has the capability to design, build and support an HAV “in the wild.” Therefore, partnerships and alliances are forming at a blistering pace between OEMs, software companies, wireless service providers, chip manufacturers, content/ecosystem providers and more.
11. While we may have imagined a robot butler serving us a drink, cars will ultimately become the first robots that have a direct impact on our lifestyle.
12. Dramatic legislative changes are coming in the next few years to deal with automated and connected vehicles, artificial intelligence and robotics, and the use of personal data of individuals, and it is clear that laws will differ widely between countries. OEMs, tech providers and intermediaries within the ecosystem must stay on top of these changes and factor the impacts into their technology builds, business models, customer documentation and contracts with counterparties. Contracts with counterparties will need to fairly allocate responsibility and risk with respect to (among other things) compliance with rules around personally identifiable information, artificial intelligence outputs and responsibility, IP ownership, cyber threats and reporting, product liability and software bugs.