Artificial Intelligence and Public Standards: report

There are currently three different sets of ethical principles intended to guide the use of AI in the public sector – the FAST SUM Principles, the OECD AI Principles, and the Data Ethics Framework. It is unclear how these work together and public bodies may be uncertain over which principles to follow. a. The public needs to understand the high level ethical principles that govern the use of AI in the public sector. The government should identify, endorse and promote these principles and outline the purpose, scope of application and respective standing of each of the three sets currently in use. b. The guidance by the Office for AI, the Government Digital Service and the Alan Turing Institute on using AI in the public sector should be made easier to use and understand, and promoted extensively.

We understand and agree that the number and variety of principles on AI may lead to confusion when AI solutions are implemented in the public sector. The principles themselves, have different histories and arose to fulfil a particular role at the international or national level. These principles fall into a small number of categories that apply in different contexts as set out below, and we have begun to signpost these to bring greater clarity.

International Norms:

The UK government has signed up to multilateral principles on AI, including the OECD principles, and is committed to implementing these through its involvement as a founding member of the Global Partnership on AI. Agreement on high-level international principles are important to establish international norms, and the principles referred to in your report are joined at the international level by the UK being a signatory to a joint statement on AI and Human Rights.

Public Sector Procurement: – both internationally and nationally

AI Procurement Guidelines – Our framework for procurement of AI in the public sector, conducted by the UK Office for AI in collaboration with the World Economic Forum and developed in consistency with the Data Ethics Framework, is another vehicle for outlining principles (drawn from OECD and FAST SUM principles) at an international level – but has also led to UK-specific AI Procurement Guide and Crown Commercial Services AI Marketplace at the national level.

Ethical Principles:

In order to ensure more clarity on ethical principles and guidance, the Government has published an online resource – the Data ethics and AI guidance landscape – with a list of various data ethics-related resources intended for use by civil and public servants. The aim is for the landing page to become a repository of data ethics papers, principles, and practical case studies. Included in this are a number of practical non-regulatory pieces of guidance as the default ways of working with AI and data across government.

• The Data Ethics Framework (launched in 2016 and most recently refreshed in 2020, widely adopted by departments and public sector)

• OAI/GDS ‘Guide for using AI in the public sector’ (how to best use AI; assess whether AI meets user needs; and ethical, fair and safe implementation of AI)

• Open Government Playbook (guidance for policymakers on being more transparent, participatory, and accountable in their work)

We acknowledge that more work is needed to keep our existing data and AI ethics guidance relevant, up to date, and user-centred, as demonstrated by our recent evidence-based process to refresh the Data Ethics Framework.

Furthermore, to bring greater cohesiveness to policy around AI and data technologies in Government, a Central Digital and Data Office has been established. On 12 January, Alex Chisholm, Chief Operating Officer for the Civil Service and Permanent Secretary for the Cabinet Office, announced the appointment of three senior Digital, Data and Technology (DDaT) leaders by the Government: Paul Willmott will Chair a new Central Digital and Data Office (CDDO) for the Government; Joanna Davinson has been appointed the Executive Director of CDDO and Tom Read has been appointed as CEO of the Government Digital Service.

The new leadership officially joined in February 2021 and are reviewing the overall digital and data programme for government. Refreshed DDaT governance structures will be considered as part of this, to ensure appropriate leadership and accountability of the government's work on data.

All public sector organisations should publish a statement on how their use of AI complies with relevant laws and regulations before they are deployed in public service delivery.

Introducing mechanisms for a more transparent use of algorithms within the government will encourage responsible public sector innovation and further enhance the UK’s long-standing leadership in the field of transparency and openness. Therefore, we will explore the development of an appropriate and effective mechanism to deliver transparency on the use of algorithms facilitating semi-autonomous decision making within the public sector. We will work in partnership with The Centre for Data Ethics and Innovation (CDEI) and other organisations to pilot the proposed approach to transparency this year, and consider what would be needed to roll it out across the public sector.

Further, the UK is a signatory to the Freedom Online Coalition joint statement on AI and Human Rights. Through this, we (along with many other countries) have agreed to a number of provisions protecting human rights, including the following:

States, and any private sector or civil society actors working with them or on their behalf, should protect human rights when procuring, developing and using AI systems in the public sector, through the adoption of processes such as due diligence and impact assessments, that are made transparent wherever possible. These processes should provide an opportunity for all stakeholders, particularly those who face disproportionate negative impacts, to provide input. AI impact assessments should, at a minimum, consider the risks to human rights posed by the use of AI systems, and be continuously evaluated before deployment and throughout the system’s lifecycle to account for unintended and/or unforeseen outcomes with respect to human rights. States need to provide an effective remedy against alleged human rights violations.

The Equality and Human Rights Commission should develop guidance in partnership with both the Alan Turing Institute and the CDEI on how public bodies should best comply with the Equality Act 2010.

In light of the request made by the CSPL last year, the Equality and Human Rights Commission (EHRC) will be developing guidance for public authorities, on how to ensure any artificial intelligence work complies with the public sector equality duty, in the next financial year. This was put on hold following initial meetings in February 2020 between the EHRC, Alan Turing Institute and CDEI. The EHRC are planning to work closely with the Alan Turing Institute and CDEI on the guidance as soon as work recommences.

Given the speed of development and implementation of AI, we recommend that there is a regulatory assurance body, which identifies gaps in the regulatory landscape and provides advice to individual regulators and government on the issues associated with AI. We do not recommend the creation of a specific AI regulator, and recommend that all existing regulators should consider and respond to the regulatory requirements and impact of the growing use of AI in the fields for which they have responsibility. The Committee endorses the government’s intention for CDEI to perform a regulatory assurance role. The government should act swiftly to clarify the overall purpose of CDEI before setting it on an independent statutory footing.

The future functions of the CDEI were consulted on in late 2020 as part of the National Data Strategy. The responses are being analysed and advice will go to ministers on the future functions and governance of the CDEI in April 2021.

The future functions set out in the National Data Strategy were: AI monitoring, partnership working, and testing potential interventions in the tech space. Whether a change to statutory status would support the CDEI to deliver its remit was also consulted on.

The AI monitoring function will provide expert-horizon scanning to identify the barriers to ethical innovation and to monitor public attitudes. Partnership working will involve working with a range of public and private sector organisations to address specific barriers to responsible innovation at an operational level, and to scale these tools and methodologies to other organisations. In addition, we are planning to ask the CDEI to build on its existing independent status to provide more practical support for the technical development of potential interventions in the tech space.

In order to identify regulatory gaps, the Competition and Markets Authority (CMA), the Information Commissioner's Office (ICO) and Ofcom have together formed a Digital Regulation Cooperation Forum (DRCF) to support regulatory coordination in digital markets and cooperation on areas of mutual importance, and enable coherent, informed and responsive regulation of the UK digital economy which serves citizens and consumers and enhances the global impact and position of the UK.

Further, the ICO, Alan Turing Institute, CDEI and Office for AI have agreed to work together to develop, roll out and monitor training for regulators on issues around AI.

Finally, the Office for AI, CDEI, and ICO and other regulators also sit on a larger Regulators and AI working group, comprising 32 regulators and other organisations. This forum will be used to discuss how to take forward the recommendations made in the report, forming a special sub-group chaired by the ICO with active membership from the CDEI, Office for AI, Alan Turing Institute, and key regulators. They will identify gaps, consider training needs and make recommendations.

Government should use its purchasing power in the market to set procurement requirements that ensure that private companies developing AI solutions for the public sector appropriately address public standards. This should be achieved by ensuring provisions for ethical standards are considered early in the procurement process and explicitly written into tenders and contractual arrangements.

During the Annual Meeting in Davos 2018, the DCMS Secretary of State signed an agreement to co-design innovative new tech policy approaches with the World Economic Forum’s Centre of the Fourth Industrial Revolution in San Francisco.

The Office for AI seconded an official to the Centre and collaborated on co-designing Guidelines for AI procurement to unlock AI adoption in the public sector. The outputs of the project aim to support innovation in ethical AI development through government procurement. The Guidelines are designed to enable government departments to accelerate the adoption of machine learning (ML) tools and ensure that the public is comfortable with the way AI-based systems make decisions. We drafted a Whitepaper with the World Economic Forum and adapted the Guidelines for the UK, publishing the AI Procurement Guidelines in June 2020 following consultation and a pilot with the Department for Business, Energy and Industrial Strategy.

The AI Procurement Guidelines recommend that procurement be carried out by diverse teams – in terms of both protected characteristics and background – in order to ensure that AI systems are representative of the diverse society they serve. The Guidelines recommend going above and beyond the legal requirements for AI services, for example, by recommending that Equality Impact Assessments be carried out in addition to Data Protection Impact Assessments, to uphold commitments in the Equalities Act 2010 and Public Sector Equality Duty.

The Guidelines support the Government’s Technology Innovation Strategy that was published by the Government Digital Service in June 2019. The Guidelines, along with the ‘Guide to Using AI in the Public Sector’ describe the foundations that each government organisation will need to best use emerging technologies including AI. They support the government’s aims and ambitions of the responsible use of AI and Data driven technologies, and guide government departments as they prepare their plans for the Spending Review.

The work on the Guidelines included early engagement with the Crown Commercial Services, who are the ultimate implementers of the Guidelines’ recommendations. In September 2020, an AI-specific Dynamic Purchasing System was launched, complementing the existing Digital Marketplace. This electronic platform provides categories such as consultancy services and ‘over-the-counter’ AI products and services. Implementing the Marketplace included establishing a baseline ethical standard that suppliers must meet to be added to the marketplace, such as the Data Ethics Framework. Crown Commercial Services engaged the market to ensure the Dynamic Purchasing System meets user needs.

The Crown Commercial Service’s Digital Marketplace The Crown Commercial Service should introduce practical tools as part of its new AI framework that help public bodies, and those delivering services to the public, find AI products and services that meet their ethical requirements.

During the Annual Meeting in Davos 2018, the DCMS Secretary of State signed an agreement to co-design innovative new tech policy approaches with the World Economic Forum’s Centre of the Fourth Industrial Revolution in San Francisco.

The Office for AI seconded an official to the Centre and collaborated on co-designing Guidelines for AI procurement to unlock AI adoption in the public sector. The outputs of the project aim to support innovation in ethical AI development through government procurement. The Guidelines are designed to enable government departments to accelerate the adoption of machine learning (ML) tools and ensure that the public is comfortable with the way AI-based systems make decisions. We drafted a Whitepaper with the World Economic Forum and adapted the Guidelines for the UK, publishing the AI Procurement Guidelines in June 2020 following consultation and a pilot with the Department for Business, Energy and Industrial Strategy.

The AI Procurement Guidelines recommend that procurement be carried out by diverse teams – in terms of both protected characteristics and background – in order to ensure that AI systems are representative of the diverse society they serve. The Guidelines recommend going above and beyond the legal requirements for AI services, for example, by recommending that Equality Impact Assessments be carried out in addition to Data Protection Impact Assessments, to uphold commitments in the Equalities Act 2010 and Public Sector Equality Duty.

The Guidelines support the Government’s Technology Innovation Strategy that was published by the Government Digital Service in June 2019. The Guidelines, along with the ‘Guide to Using AI in the Public Sector’ describe the foundations that each government organisation will need to best use emerging technologies including AI. They support the government’s aims and ambitions of the responsible use of AI and Data driven technologies, and guide government departments as they prepare their plans for the Spending Review.

The work on the Guidelines included early engagement with the Crown Commercial Services, who are the ultimate implementers of the Guidelines’ recommendations. In September 2020, an AI-specific Dynamic Purchasing System was launched, complementing the existing Digital Marketplace. This electronic platform provides categories such as consultancy services and ‘over-the-counter’ AI products and services. Implementing the Marketplace included establishing a baseline ethical standard that suppliers must meet to be added to the marketplace, such as the Data Ethics Framework. Crown Commercial Services engaged the market to ensure the Dynamic Purchasing System meets user needs.

Government should consider how an AI impact assessment requirement could be integrated into existing processes to evaluate the potential effects of AI on public standards. Such assessments should be mandatory and should be published.

The updated Data Ethics Framework contains a set of questions that should be answered throughout any data science or AI-related project. Questions help teams understand the impact of their projects and the simple scoring system flags the need for additional work and guidance when improvement of ethical standards in the project is needed.

Further, the Office for Artificial Intelligence is currently working on a National AI Strategy, based on recommendations from the independent AI Council AI Roadmap published in January as well as consultation with the AI ecosystem across a range of issues, including ethics and unanticipated impacts of AI. The Strategy aims to be both globally ambitious and socially inclusive, and will publish later this year.

Government should establish guidelines for public bodies about the declaration and disclosure of their AI systems. Recommendations to front-line providers, both public and private, of public services The Committee makes seven recommendations to front-line providers of public services to help establish effective risk-based governance for the use of AI.

Ensuring fairness in how the public sector uses algorithms in decision-making is crucial for gaining and maintaining public trust. Introducing mechanisms for a more transparent use of algorithms within the government will encourage responsible public sector innovation and further enhance the UK’s long-standing leadership in the field of transparency and openness. Therefore, we made the following commitment in the National Data Strategy published in September 2020:

We will collaborate with the leading organisations and academic bodies in the field to scope and pilot methods to enhance algorithmic transparency.

We are currently working with the Centre for Data Ethics and Innovation and the leading organisations in the field to develop and pilot algorithmic transparency measures.

As part of the work on the new National AI Strategy, to be published later this year, the Office for AI will be building on previous work conducted as part of the Guide for AI in the Public Sector and AI Procurement Guidelines towards ensuring use of AI in public services is transparent and works for the benefit of citizens.

Providers of public services, both public and private, should assess the potential impact of a proposed AI system on public standards at project design stage, and ensure that the design of the system mitigates any standards risks identified. Standards review will need to occur every time a substantial change to the design of an AI system is made.

Providers of public services, both public and private, must consciously tackle issues of bias and discrimination by ensuring they have taken into account a diverse range of behaviours, backgrounds and points of view. They must take into account the full range of diversity of the population and provide a fair and effective service.

Providers of public services, both public and private, should ensure that responsibility for AI systems is clearly allocated and documented, and that operators of AI systems are able to exercise their responsibility in a meaningful way.

Providers of public services, both public and private, should monitor and evaluate their AI systems to ensure they always operate as intended.

Providers of public services, both public and private, should set oversight mechanisms that allow for their AI systems to be properly scrutinised.

Providers of public services, both public and private, must always inform citizens of their right and method of appeal against automated and AI‑assisted decisions.

Providers of public services, both public and private, should ensure their employees working with AI systems undergo continuous training and education.