15 May 2023

Increasing column inches are being filled in relation to Artificial Intelligence (AI) and the potential impact on many sectors, demonstrating evidence of growing excitement and concern in almost equal measure. You will have heard the warning from Geoffrey Hinton, previously of Google, that AI chatbots could soon be more intelligent than humans, largely due to the speed of advances which are approaching an exponential rate. There have been calls to pause progress in such developments in order to design and implement robust safety measures.

How then, might such rapid development impact upon the framework for Automated Vehicles (AVs) and their potential to become more commonplace?

Starting with the present position, it seems there is still some way to go before we can step into a driverless vehicle and allow it to take us where we want, however much that might appear to be the commonly promoted image of what an AV is in social media and television advertising. The small print often states such images are “conceptual” rather than demonstrating actual developments on the production line. Nevertheless, the attraction of AVs is largely that of safety – if we can eliminate human error from driving, then there is an obvious and immediate benefit to all.   However, can we be sure that the technology is reliable and that it won’t fail us at a critical moment? Are we willing to trust our safety to it?

AVs will still require human drivers to be responsible for taking back control when need be. The Law Commission of England & Wales and The Scottish Law Commission prepared a joint report in 2022 pointing out that there is a division between legal rules governing vehicle design and those covering driver behaviour. Ultimately, the Commissions made 75 recommendations, aiming to strike the balance between regulation and innovation in order to allow progress to continue and reap the benefits, while mitigating the risks.  Tests and trials continue to be rolled out, with a trial over the Forth Road Bridge crossing, from Fife to Edinburgh Park, due to commence later this month.

There is a conceptual leap to be made, as, although the driver of an AV would benefit from the technological capabilities, they have to remain fully responsible for the driving task. It is anticipated that the AV would make a request to the driver to resume control. This is referred to as the Human Machine Interface (HMI), but how long would be required for a driver to regain situational awareness and take back control? Estimates in studies range from 8 seconds to 40 seconds.

Is a driver in an AV likely to be quite as alert as one driving a normal vehicle?   In the California State Court April 2023 decision of Justine Hsu v Tesla, a driver was unsuccessful in suing the car company when the car crashed while she was using the autopilot function. She was found to have been using the function incorrectly with the manual clearly stating that autopilot should not be used in city streets.

There would appear to be a tension between the messaging from the car manufacturers and driver expectation, therefore. Drivers are likely to be expected to read, understand and accept lengthy terms and conditions, and to follow manufacturer’s instructions to the letter. Such requirements are not dissimilar to those faced when browsing the internet or signing a contract for financial services, and we know that people do not always read the full T&Cs in such circumstances.

Where their personal safety and that of other road users is at stake, will people be more careful to read and understand what they are signing up to?

Perhaps there is a way to go in seeking to narrow the gap between vehicle design capability and driver behaviour identified by the Law Commissions last year. We await the outcome of the next set of trials with interest.

Tim Webster, Associate: twe@bto.co.uk / 0141 221 8012