In response to a recent whitepaper consultation on Artificial Intelligence (AI) regulation, the Chartered Institution of Highways & Transportation (CIHT) has advised the UK Government that there needs to be a greater consideration of the role data, local authorities and national bodies have in supporting AI technologies.
The CHIT has said there needs to be a greater consideration of the role data plays in supporting AI technologies, and this should be reflected in the cross-sectoral AI principles. It also said local authorities and national bodies will be key to rolling out AI in public services such as transport, and so should be given appropriate funding, guidance, and procurement frameworks to do this successfully.
The CIHT called for a ‘clear evidence-based approach’ to policy developments is critical, particularly when it comes to public understanding around the adoption of new and emerging technologies within the transport sector. Regulators and organisations such as CIHT should work together to ensure that unbiased evidence on the pros and cons of AI is well communicated and shared widely. Working across the sector to inform and educate people will build a healthy relationship between users and AI, it said.
The Government is now consulting on their recently published AI regulation whitepaper, which seeks to progress the AI capability of the UK by bringing clarity and coherence to the regulatory landscape.
The whitepaper has put forward five cross-sector principles that will be issued on a non-statutory basis and implemented by existing sector regulators, to ensure that all existing and developing AI technologies are safe, fair and reliable.
These cross-sectoral principles are:
- Safety, security and robustness
Regulators will offer guidance on cybersecurity practices and considerations of privacy practices. They will refer to risk management frameworks and ensure that AI systems are technically secure and function reliably as intended throughout their entire life cycle.
- Appropriate transparency and explainability
Regulators will establish expectations for organisations and individuals that deploy or operate AI to provide information on the AI’s nature, purpose, data, logic, and accountability. Regulators will also set explainability requirements, particularly for higher-risk systems, balancing regulatory needs and technical trade-offs.
Regulators should interpret and articulate fairness in their sectors, determining when it is relevant. They must establish governance requirements for fairness, especially when AI decisions have significant effects on individuals.
- Accountability and governance
Regulators will determine who is accountable for compliance with existing regulation and provide guidance on how to demonstrate accountability.
- Contestability and redress
Regulators will develop or revise guidance on how to address complaints or disputes related to AI harms and specify formal avenues for redress.
In response to the whitepaper, CIHT highlighted that these cross-sectoral principles are lacking in their consideration of the importance of data and how it is collected. AI technology can only be as good as the data it is relying on, as this is all the AI system knows about the world. Therefore, without complete and comprehensive data sets none of the current cross-sectoral AI principles can be met.
CIHT has called for the AI regulation framework to ensure that the data feeding systems is:
- fit for purpose, recorded in standardised formats on modern, secure, future-proof systems.
- held in a condition that means it is findable, accessible, interoperable, and reusable, and accords with open data standards where possible.
When it comes to the transport sector, local authorities and national bodies are key to delivering, managing, and maintaining much of our infrastructure that enables the movement of people and goods. Because of this, it will be vital that any national advancements in technologies such as AI, ensure that the capabilities of local authorities are not left behind.
Local authorities are currently facing significant fiscal constraints: inflation and spiralling energy costs have impacted budgets considerably against remaining pressure to ensure the delivery of statutory services. Thus, many may well not have the financial and staffing resources to invest in the technology necessary to enable AI use within their organisations. Local authorities will therefore need specific funding to invest in AI technologies, guidance to support the delivery of AI, and procurement advice to contract these technologies effectively, said the CIHT.
“Scepticism towards new technologies is not new within the transport sector, as we have seen recently with smart motorway projects and autonomous vehicles. In both these examples, the primary concerns of the general public often focused on the reliability and safety of these new technologies. It is likely that any future AI technology that involves controlling or operating moving vehicles will likely receive the same concerns, and so a clear evidence-based approach to developing and regulating AI in this sector will be critical,” it added.
“It is important that the factual evidence of the safety and reliability of AI is not misreported in the media or by those who promote such technologies, and that public trust towards AI is built up by being honest about its failures and how to avoid this whilst promoting its benefits and how they can be harnessed in all sectors.”