show image

Lord Clement-Jones

Co-Chair of the All Party Parliamentary Group on AI

12 tech policy priorities for 2020

We don’t know at this stage exactly how the structure of the Department for Digital, Culture, Media and Sport is going to change with the reshuffle at the end of this month but since we have recently had the twelve days of Christmas I recently used my very brief few minutes during the Queen’s Speech debate in the Lords to suggest 12 digital priorities for action by this new government.

They are a mixture of priorities which impact variously on consumer, citizens and business stakeholders. We need to retain public trust and increase public understanding in these digital technologies, which means we need an active government digital policy to reflect the new protections needed but building on what so far has been achieved in the UK.

1. Action on online harms

The big question here is whether the government will bring forward an early draft bill for scrutiny so that definitions of harms and the scope of platforms affected can be debated and we can discover whether it will designate Ofcom as the regulator to ensure the necessary work on codes of practice are put in place.

Online harms should also cover the use of social media in elections and its impact on our democracy. The government is yet to adopt the recommendations as to transparency of source and funding by the Electoral Commission in its Digital Campaigning: Increasing Transparency for Voters report of June 2018 which called for urgent reforms to electoral law to combat misinformation, misuse of personal data and overseas interference in elections. Both the Democracy Disrupted report by the Information Commissioner into the activities of Cambridge Analytica and the report of the intelligence and Security and Intelligence Committee on Russia are relevant in this context.

2. Speed up broadband rollout

I welcome the government’s pledge to accelerate the rollout out of gigabit capable broadband. Whether this matches the prime minister’s pledge of 100 per cent fibre to the home by 2025 during his leadership campaign however is unclear, as is the relationship to the rollout of 5G or whether changes to the current extraordinarily low Universal Service Obligation of 10mps are contemplated.

3. A moratorium on facial recognition technology

Generally my preference is for ethical codes for the regulation of new technologies but there are some technologies such as live facial recognition technology which need urgent regulation. This was revealed to have been in use for some time in the Kings Cross area of London, was the subject of a High Court decision in Wales and has been described as potentially Orwellian by the Metropolitan Police Commissioner and deeply concerning by the Information Commissioner.

The Home Office’s own Biometrics and Forensics Ethics Group has questioned the accuracy of live facial recognition technology and noted its potential for biased decision-making.

A moratorium as recommended by the Science and Technology Select Committee is a vital first step. We need to put a stop to this unregulated invasion of our privacy and I have tabled a private members bill on the subject.

4. Control of algorithmic decision-making

Even more prevalent than facial recognition is the use of algorithms in decision-making in the public sector. One third of councils are now using algorithms to make decisions about benefit claims and other welfare issues. They are now being rapidly adopted in policing and probation too.

The government has published a guide to using artificial intelligence in the public sector and and signed up to the WEF procurement rules for AI. But what does that signify if there is no central body tasked with monitoring their use and ensuring compliance in the delivery of decisions and services whether in health, justice, policing or otherwise?

This is particularly important as regards explaining decisions and ensuring they are free from bias. As Google’s Andrew Moore says: “The era of black box machine learning is behind us”.

The UN Rapporteur on extreme poverty and human rights has also expressed concern about the emergence of the digital welfare state. The question of when the government is going to properly regulate algorithmic decision-making is something we should all be pursuing in this parliament.

5. Improved data governance

Another matter of growing concern is the tech industry’s handling of one of our most rapidly growing resources, our data. Our individual personal data, in almost every case, is the currency of these new technologies and the means of training intelligent software.

Though personal data rights have been strengthened by the introduction of the General Data Protection Regulation (GDPR), the public is increasingly concerned about how they can manage who has access to their data and what they are able to do with it, while at the same time enjoying the services that they receive.

This includes the NHS where we still have had in recent times reports of inappropriate data sharing.

We need to go well beyond the the GDPR with much more control over our personal information through personal data accounts and data trusts. We need to ensure interoperability and that value attributed to our data so this isn’t just something that can be collected without its value to individuals or society being recognised.

The Open Data Institute, Alan Turing Institute and others are doing good work on this and I hope the government will continue to fund their work.

6. Changes to corporate governance

Artificial intelligence in particular has huge implications for the future of Corporate Governance. I welcome the fact that as part of the announced Workers and Families Employment Bill there will be proposals on company audit and corporate reporting including a stronger regulator designed to improve public trust in business.

There is a growing corpus of corporate governance work relating to AI and the ethics of its application in business. Asset managers are now adopting guidance for the companies they invest in.

As I said in my last piece for NS Tech, companies need to ingrain ethical behavior. AI can and should contribute positively to a purposeful form of capitalism which is not simply the pursuit of profit but where companies deploy AI to achieve greater sustainability and a fairer distribution of power and wealth, not exclusively driven by returns to shareholders. It is imperative that boards have the right skill-sets in order to fulfill their oversight role and take accountability.

They must be aware of the questions they should ask and the advice they need and from whom. They need to consider and report on what tools they need and use.

The question is whether the government will ensure that the Finance Reporting Council or the Audit, Reporting and Governance Authority in its new incarnation will take this forward in a new corporate governance code for the AI age as part of these proposals on reform of company audit and corporate reporting.

7. A new approach to skills, retraining and diversity

AI in particular will have significant implications for the ways in which society lives and works and will accelerate the disruption of the jobs market. In terms of education, skills and reskilling, this means we must ensure we are prepared for the future automation of many jobs. Whatever the scale of disruption, retraining will be a lifelong necessity.

However, despite this, the government National Retraining Scheme is still only at pilot stage. The expected pace of rollout is currently unclear as is its relation to the £3bn investment in the new National Skills Fund.

In terms of the skills we should be nurturing, it is very clear that this is not just tech skills such as maths and coding. Tech developers agree that social and creative skills and critical thinking will be needed as well – so the humanities will be as important as the sciences in our schools. We need to teach the core skills required for children to flourish in the modern world, including critical thinking, verbal reasoning and creativity.

The Future Now initiative recently launched by the ex Lord Mayor Sir Peter Estlin recognises that there are some 17.3 million people in the UK without the necessary digital skills for work but that 90 per cent of UK jobs within 20 years will require digital skills. This raises the question of what progress has been made through Apprenticeship Levy to the creating apprenticeships in the tech sector.

There is also crucially the importance of ensuring that the data being used for training testing or operational inputting does not exhibit bias. We must not build in the biases and prejudices of the past in the datasets we use and the algorithms we use. As a result it is really important that we have the necessary diversity/inclusion in the AI workforce to spot problems of bias in training data and decision making. The irony in terms of gender balance in the tech space is that we have been going backwards since the Second War and have only recently started to address the issue. The government needs to be taking steps to encourage greater diversity in the training and recruitment of AI specialists.

8. Strengthening of digital competition

It is clear we need new competition tools to level the playing field where data is concerned so that our SMEs are not hindered from innovating through lack of access to training data. The creation of a Digital Markets Unit as recommended by the Furman Report into Unlocking Digital Competition will be a significant step.

Competition regulation for the digital industries needs to be substantially strengthened so that the importance of data as an asset is recognised and we prevent data monopolies that form barriers to innovation.

9. Adoption of international norms on ethics

There are choices to be made about the uses to which AI and other technology is put. Professor Margaret Boden expresses it very simply: “Even if AI can do something, should it?”

We must seek to actively shape AI’s development and utilisation, or risk passively acquiescing to its many likely consequences. The cardinal principle is that AI must be used for the benefit of society. Does it better connect and empower our citizens? Improve working life? Create a more sustainable society?

Whatever good work we are doing at the national level, we need clear international agreement on the ethical development of AI. The OECD recommended guidelines for ethical AI containing five AI Principles were adopted by OECD members and others (42 countries) in May last year. But the UK needs not only to sign up to but also develop mechanisms for compliance with international and global standards to prevent an ethical race to the bottom and so-called AI nationalism. This includes moves towards a convention designed to control Lethal Autonomous Weapon Systems.

10. Reform of digital taxation

The government plans to introduce a Digital Services Tax ensuring that the tech giants pay their fair share. I can understand this as a short-term response but I hope we won’t be going it alone. We need to place emphasis on supporting and building on the OECD’s proposals which could have international application and be much more digital trade friendly.

11. Ensuring better media literacy

As a major priority we also need to make sure, as Doteveryone, 5Rights and others emphasise, that we are all better prepared for working with AI and navigating an AI-driven world in terms of general digital understanding and literacy, allowing us not simply to be digital victims but empowered to make the best personal choices using these new technologies.

At minimum, we need to include teaching in the curriculum about how to use social media responsibly and provide advice and support for parents on how to help their children protect themselves online.

12. A strong digital trade policy

At the end of this month, we will need to begin realising an international trade strategy, negotiating the future relationship with the EU and US among others. Our thriving tech sector could be profoundly affected by the wrong decisions. I hope our new government will put digital trade at the heart of its trade policy.

For instance without a data adequacy agreement with the EU, will we be unable to manage data flows between the UK and our European partners, meaning that the UK’s digitally intensive sectors, which account for 24 per cent of total UK exports, will experience a major disruption. We have the prospect of having a lack of tech skills from the continent. There are major other issues which will emerge in any trade negotiations with the US and EU not least our position on data sovereignty and as regards the US Cloud Act and the EU Digital Services Act.

So there is plenty of work to do in the digital area. Let’s hope a government with a majority of 80 has the appetite to tackle some of these issues.

Tim, Lord Clement-Jones is the former Chair of the House of Lords Select Committee on AI and Co-Chair of the All Party Parliamentary Group on AI.