Thanks to all our panel and everyone who sent in questions
The chat has ended. Many thanks to our panel and readers for taking the time to join us today. We hope you got as much out of this discussion as we did and we look forward to exploring the points raised in further depth over the coming weeks.
Updated
What should the priorities be?
To finish, I’d like to ask each of the panelists to give their views on what the government’s priorities should be when it comes to regulating autonomous vehicles and their data?
With safety as the top priority, new regulations are necessary. At the same time, there are opportunities to facilitate development and testing for companies
developing these solutions. Safe, robust AI algorithms are improved through the collection of large amounts of data. Ideally, we would be able to test fleets across all states with their diverse driving conditions. A patchwork of different regulations in different regions hampers development and progress. It would be enormously beneficial to have a unified set of regulations across various regions and countries.
A clear regulatory process for gaining permission / license to operate fully autonomous (L4/5) service should be the Govs priority. This clarity will enable private enterprise to build business which can deliver on the promise of AVs.
A clear and robust safety validation process should be developed by industry and approved by Gov. Gov intervention may be required to ensure comprehensive access to hazard scenarios and test cases which will be essential for delivering safer systems and thus gaining public trust.
The Government might consider a mechanism for stipulating that autonomous vehicle programs testing or deploying in the UK share the real world hazard events that they encounter so long as making such data widely available does not diminish the UK’s ability to build companies with first mover advantage in its key markets.
Updated
Panelist Birgitte Andersen raises an important issue
The issue of autonomous decisions and control is a huge issue. How does the car make a selection decision if it is choosing between two people in an accident – and how will it decide between a dog and a child?
Keeping our data safe
Inevitably, when we talk about data, we need to talk about cybersecurity – what measures will need to be taken to keep that data safe? Whose responsibility is this?
The Government recently called on Manufacturers to up their game in this area, but there was no real detail. Insurers wont be allowed to exclude Cyber risks as we need the public to feel safe and be looked after if there is an issue. Within the consortia we are working with there is a good focus, Flourish is looking specifically at that area, and within Venturer we have BAE Systems as a Partner, and they do military grade Cyber Security as a business, so good people to have on board.
International and national co-operation on standards is key and much good work is ongoing with BSI and others directly involved in the thinking. One of the other issues that Government will need to consider is whether CAV operators will be designated as "operators of essential services" under the Network and Information Services (NIS) Directive. That would increase the level of action required but add complexity. There is also of course a Brexit aspect to that decision.
My nightmare scenario is that your autonomous car is driving along some small country road in the middle of nowhere. It could be your car or even one which is part of some new non-ownership scheme. It all of a sudden pulls over, everything shuts down apart from the screen which now reads - “We’ve hacked your car! Pay 100 bitcoin to get it back.” (more like 0.5 bitcoin at the current exchange rate)
We need to work hard to try ensure this doesnt happen.
Updated
A question of ethics
I am internally challenged by wanting to know that data is being used to protect me, to inform me and to improve my journey experience, versus not wanting to be controlled, policed and manipulated into a behaviour - what are the panels thoughts on my dilemma?
Guardian contributor I think we need a set of digital ethics to help us navigate this uncharted factor. The data will be there, and protocols for anonymising, safe keeping, using and disposal will be essential, as well as making sure its used for the right and appropriate reasons.
There is an argument that data relating to hazard events / near misses, etc. that were learnt during real-world testing should be shared publicly thus to improve the safety performance of all technology vendors: i.e. the public were put at some risk therefore details should be made available for public good
The need for data analysis
As a design consultant we may need to use the services/ support of partners that have prior / proven experience in data analytics to be able to glean insights. Is that the approach we may be taking?
The amount of data will be staggering and much of it will be incredibly insightful. We are looking at partnering in many ways, but first we need to establish who will have access to that data. If most of it only goes to the Motor Manufacturer, that could stifle a vibrant potential economy driven from third party entrepenurial ideas and applications. That debate will get quite lively I think!
Updated
What will the impact be on infrastructure when it comes to driverless cars?
Guardian contributor Infrastructure and the built environment will absolutely be impacted, and we hope positively. These data types can inform the design of roads, car parks, garaging, charging sites etc, as we will know more about how these vehicles will interact with the built environment, and may change how we interact with it. Lets also remember that this data can inform the three layers of Infrastructure: Physical; Digital and Commercial.
Updated
Should we deregulate in the UK to remain competitive?
A great question from the floor:
Would the panel agree that in order for the UK to remain competitive in the CAV space, and encourage inward investment, the government must be prepared to push the boundaries of real-world testing and regulate/de-regulate accordingly?
Our panel is broadly in favour:
There is clearly a race to be first in the driverless car market – huge ‘engine’ for growth for a nation or company to dominate the market.
I agree on inward investment and pushing boundaries. We must not forget that for CAV to work the huge private investment of companies and those buying the cars have to be matched with a huge public infrastructure investment. It is not just about perfect technology, or ethics rules regarding controls and automated decisions or data governance.
Infrastructure investment is about investement sensors to work in conjunction with an integrated transport network. It is also about investment in a charging and powering infrastructure to the cars. And much more..
Clarity on how operators and technology vendors can achieve regulatory approval to put these technologies in to action on Britain's roads will be what has and will continue to encourage inward investment. But our focus shouldn't be entirely on real world testing. It's important, but virtual validation methods are equally important. In simulated worlds we can drive many more miles, and we can make each and every one interesting and critical in some way to exploring potential failure modes of these systems. It's also something that the UK should be good at, given our pedigree in the adjacent technology space of computer gaming
I think the Government is pushing boundaries to a degree and is ahead of most other countries with regard to regulation. It needs to be balanced however, Human Safety is paramount, and an early accident due to an over zealous approach could derail the whole programme.
Updated
Getting to grips with the data
To kick us off, let’s understand the scale of the issue. What sort of data will be collected by autonomous cars and what provisions need to be made to manage it?
Guardian contributor Good question to start with! At this stage (testing and trialling of an individual vehicle) the sort of data being collected is - journey data, vehicle performance data, user behaviour data, booking data, disruptions data, deviations data, environment data, sensor data (from roadside infrastructure) as well as GPS data to support journey deviations. Probably more types also! At the moment its being managed and controlled in certified test sites and servers, but obviously this cant sustain a more widespread CAV network. Server Space and 5G connectivity are two of the grand challenges for CAV implementation.
Data that will be collected in the future from CAVs will fall into four main categories: non-sensitive data (such as congestion information and anonymised information); personal data (such as location or biometrics - if used) ; special category data (for example collision data and anything relating to health condition) and commercially sensitive data (for example manufacturer data relating to its own IP). The arrangements in place in the UK/Europe will need - from May next year to comply with the updated consumer and other protections in the EU Data Protection Regulation.
Updated
The chat has started
Welcome to today’s live chat. In the next 90 minutes, our panel will be discussing how to manage and regulate the data generated by autonomous vehicles as they become more commonplace on Britain’s roads.
Join us as we debate the legislative, ethical and administrative questions involved, and ask who does the data belong to? Please do put your questions to the panel in the comment space below.
Updated
Some background reading to kick off our debate this afternoon
Updated
How to join in the discussion
Make sure you are a registered user of the Guardian (if not, it’s quick to register) and join us in the comments section below on 7 September, from 12:30-2pm.
You can send also questions for the panel in advance by emailing emma.sheppard.casual@theguardian.com or by tweeting @GuardianPublic using the hashtag #CAVdatachat.
Updated
The panel
- Chris Jackson, head of transport, Burges Salmon
- Justyna Zander, technology director, NVidia Corporation
-
David Williams, technical director, AxA
- Birgitte Andersen, CEO, The Big Innovation Centre
- Ben Peters, co-founder, FiveAI
- Melanie Smallman, deputy director Responsible Research and Innovation hub, UCL
- Nathan Marsh, UK & Northern Europe Director, Intelligent Mobility, Atkins
- William Sachiti, founder, Academy of Robotics
Updated
What we’ll be discussing
The rapid advancement in technology means up to 25% of new vehicles sold in 2035 could be fully autonomous, according to a recent report. In 2016, the government supported plans to begin testing on motorways as early as this year. But as connected and autonomous vehicles (CAVs) move on from pilot stage to becoming a reality on Britain’s roads, it’s crucial for policy makers to be clear about how they will be regulated.
The amount of data that will be produced by each autonomous vehicle is staggering. Intel estimates 40 terabytes will be consumed and generated for every eight hours of driving, and a million cars will generate the equivalent of three billion people’s data. That’s thanks to the onboard sensors and cameras, detailed maps and technical information about roads and hazards, which have to be constantly updated. Personal data will also be stored.
But in a world where all journeys will be traced and identified, who should have access to that data – does it belong to the companies developing the vehicles, or the driver? How will this information be controlled, monitored and regulated? What are the ethics surrounding data collection? What consequences will there be for infrastructure and urban planning?
Join us and our panel of experts in the comments section of this page on Thursday 7 September from 12:30-2pm (BST), to discuss these questions and more.
Updated
Interesting discussion all! In terms of government priorities, I think that the focus should be on firstly ensuring that autonomous vehicles are helping build the kind of kind of transport system we want in the UK - these vehicles are going to have a profound impact on the way we travel and possibly the shape of our cities and this shouldn't be seen as inevitable. Secondly the government needs to look at data ownership and privacy, with the aim of making sure that any policy reflects the values and concerns of the British public.