Centre Write

Every new technological revolution brings with it the same existential question about the future of work: is a robot going to take my job? In the past, it was the relatively routine work in agriculture, manufacturing and the service sector that got replaced by physical machines, robots and computers. 

But today, well-paid roles in finance, tech and publishing are at most risk of disruption, and not from physical machines, but digital ones: artificial and remote intelligence.

Advances in artificial intelligence (AI) capabilities are now exponential, and it is not just their pace that we need to be aware of, but also the breadth of their application. 

Many of these developments are to be welcomed. They have the potential to make the world of work more productive. Teachers are using AI to create work plans and personalise learning, lawyers are piloting it to automate and enhance contract analysis, due diligence and regulatory compliance and developers are using it to auto-generate new software code.

But there are concerns too. Millions of professional service jobs could be at risk of automation. GPT4 can learn to trick humans, making online fraud and harassment easier and more widespread, autonomous AI bots can be used to undermine national security, and this is all before we get to the potential for super-intelligent, God-like AI that poses catastrophic risks for humanity.

Fifty years ago, the development of the first silicon microchips transformed computing power and ultimately kickstarted the ICT revolution, leading to rapid deindustrialisation through the 1970s and 1980s. While we were quick to embrace the productivity power of new technologies, we did not do enough to help individuals and their communities adapt to the new realities – that meant manual and clerical jobs were automated or offshored overnight. 

Half a century on, recent developments in AI suggest this wave of technology may be different, as AI systems are capable of automating the non-routine tasks of white-collar workers. The UK’s high-productivity services sector could be threatened by AI, with the potential to push down the global cost of these services by making it easier and quicker to provide them by competitors across the world. But this time, we have the chance to learn from our past mistakes.

Goldman Sachs recently forecast that around 24 million jobs in the UK will be exposed to automation, and that 25% of work tasks could be automated with existing AI technology. Other studies suggest jobs requiring a bachelor’s degree could be most affected, with over 30% of them expected to have at least 50% of their tasks exposed to large language models such as GPT4. 

Indeed, AI is already disrupting the jobs market. It is being used to determine gig workers’ pay and hours, and to hire, monitor and manage office workers. BT and IBM are among the companies that have announced that tens of thousands of employee roles will be replaced by AI, with BT planning to replace 10,000 roles and IBM planning to replace 7,800. Examples of highly exposed jobs include information services, finance, publishing and telecommunications. We are highly dependent on services and creative exports — areas already being disrupted by generative AI and easily automatable under the even more sophisticated artificial general intelligence (AGI).

To ensure AI is a tool to support workers and not a threat to their jobs, the Government must focus on three areas. 

First, it must produce regular reports – potentially powered by AI – forecasting the expected impact of AI on the domestic and global jobs market. 

Second, an AI training fund is needed to incentivise business and support people to learn to use AI as a productivity tool in their daily jobs or help them retrain for new roles. This training should be delivered by AI-enabled personalised learning and a digital learner ID that links formal and informal qualifications. 

Third, the Government ought to introduce a new era of digital workers’ rights that enable businesses and workers to use tech to support their wellbeing and productivity. 

It is easy to dismiss new tech as a passing fad: web3, NFTs and crypto each were the future once. However, AI has the potential to be radically different, affecting everything, everywhere, all at once. The tech revolution is happening to our economy, to our jobs and to our politics, whether we are prepared for it or not. We can benefit from it, but only if we act with intent.

Jeegar Kakkad is the Director of Future of Britain Policy at the Tony Blair Institute for Global Change

This article was published in the latest edition of Centre Write. Views expressed in this article are those of the author, and not necessarily those of Bright Blue. 

Read more from our August 2023 Centre Write magazine, ‘Back to business?’ here.

Taken together, the two ways through which the Government is attempting to resurrect its flailing asylum agreement with Rwanda – a new treaty and emergency legislation – are a serious threat to the human rights and safety of refugees, the UK’s commitment to the European Convention on Human Rights (ECHR) and the independence of the British judiciary.

The Government’s plans follow from the UK Supreme Court’s conclusion that, after looking impartially at the evidence, Rwanda is not a safe country to send asylum seekers to. The Supreme Court found that Rwanda has a long history of misunderstanding its obligations to refugees and does not comply with those obligations effectively – in some instances, even wilfully disregarding them. Therefore, there was a real risk that people could face serious harm in Rwanda. The Government’s plans do not satisfactorily address these detailed and rigorous findings and are a blatant attempt to prohibit any further independent judicial assessment of this risky policy.

Starting with the treaty, Ministers and officials have tried their best to change the content and wording of the agreement in light of the Supreme Court’s concerns. Paragraph 85 of the Supreme Court’s judgment, for example, highlighted the 100% rejection rate of asylum claims from Afghanistan, Syria and Yemen in 2020-22 – based on Rwanda’s view that asylum seekers from the Middle East should remain in the region. In response, Article 3(1) of the treaty requires decisions to be made on a non-discriminatory basis without regard to nationality.

Further, paragraph 87 of the judgment evidenced Rwanda’s practice of refoulement – that is, returning people to countries where they could be seriously harmed or killed – up to the present day. Indeed, during the negotiations for the original deal, six people were unlawfully returned to unsafe countries. In response to these concerns, Article 10(3) of the treaty states that nobody may be removed from Rwanda, except to the UK, even if their asylum application is unsuccessful. 

Moreover, based on the Supreme Court’s concerns about the independence of the Rwandan judiciary in the appeals process, Annex B paragraph 4.2 of the treaty establishes a separate Appeal Body which requires a Rwandan and Commonwealth president and co-president who will appoint judges to sit on the Body. 

As far as words go, these are good ones. But the point of the Supreme Court’s decision was never the words. The point was whether Rwanda could be relied on to understand and adhere to those words. Looking at the evidence provided by all parties, including the Government, the Supreme Court concluded that the answer was no. Paragraph 88, for example, found that Rwanda had “a culture of, at best, insufficient appreciation by … officials of Rwanda’s obligations under the Refugee Convention, and at worst a deliberate disregard for those obligations.” 

This evidence has not changed since the Supreme Court’s judgment. Therefore, the commitments set out in this treaty do not resolve the most important question of whether Rwanda will in fact comply with its obligations. 

What is more, the Supreme Court was clear that making Rwanda safe would require structural and cultural change within Rwanda’s judicial system. For example, paragraph 105 states that:

“We accept … that the capacity of the Rwandan system (in the sense of its ability to produce  accurate and fair decisions) can and will be built up. Nevertheless, asking ourselves whether  there were substantial grounds for believing that a real risk of refoulement existed at the  relevant time, we have concluded that there were. The structural changes and capacity-  building needed to eliminate that risk may be delivered in the future, but they were not shown to be in place at the time when the lawfulness of the policy had to be considered in  these proceedings.”

It is highly unlikely that these changes have happened in less than one month – meaning that there remains a real risk that people will be seriously harmed by this policy.

Making matters worse, the Government’s emergency legislation is a blatant attempt to prevent any further independent judicial scrutiny of its policy, which guts the effective protection of basic human rights and threatens the UK’s commitment to the integrity of the European Court of Human Rights.

Clause 2 of the Bill, for example, requires all decision makers to conclusively treat Rwanda as a safe country, despite the Supreme Court’s decision. In effect, this tries to put a fiction into British law – a fiction that will put innocent people’s lives and safety at risk. Worse, Clause 2(3)-(4) bans judges from considering an appeal or judicial review on the basis that Rwanda is an unsafe country.

Clause 3 continues this regressive trend by disapplying much of the 1998 Human Rights Act 1998 (HRA). Judges will not, for example, be allowed to interpret the legislation so that it respects the ECHR (Clause 3(4)), and the Home Office will not have a legal obligation to respect the ECHR when making decisions about removing people to Rwanda (Clause 3(5)).

Instead, the only safeguard against deportation to Rwanda  is for individuals to argue based on “compelling evidence” that their particular circumstances create a “real, imminent and foreseeable risk of serious and irreversible harm.” This is a very high threshold – and with the difficulty accessing legal advisers in immigration and asylum – the result is that many people will struggle to successfully launch a challenge and face real risks of serious harm.

Finally, Clause 5 requires judges to ignore any interim measures of the European Court of Human Rights and, instead, gives an absolute right to the Government to decide whether the UK will adhere to them. An interim measure is an order requiring the Government to temporarily do or not do something to secure the protection of human rights. Interim measures are granted only exceptionally when individuals face a real risk of serious and irreversible harm. For example, in 2022 two British citizens – Aiden Aslin and Sean Pinner – secured an interim measure against Russia ordering Russian separatists in occupied Ukraine to execute them.

This Clause undermines the independence of British judges and weakens the UK’s commitment to the integrity of the European Court of Human Rights. A mature democracy committed to the rule of law should not compel its judges to ignore orders designed to secure people’s human rights.

It is little wonder that the Home Secretary is unable to certify that the Bill is compatible with the ECHR – indeed, the Bill is specifically crafted to remove and weaken the ECHR’s protections.

The whole Rwanda saga has been a sorry episode from start to finish.  Expensive – costing £240m so far, with £50m extra expected – unethical and unlawful, the policy has been an exemplar of poor decision-making that will be taught for years to come. Core principles of which British governments should be proud champions  – human rights, the rule of law, judicial independence and the liberal international order – have been critically undermined and with nothing positive to show for it. The world’s most persecuted refugees, the British people and the UK’s principles deserve better than this.

Lee Marsons is a Senior Researcher at the Public Law Project.

Views expressed in this article are those of the author, and not necessarily those of Bright Blue. 

There is no doubt whatsoever that big tech must be regulated. US legislators say it; EU authorities agree; the UK body politic concurs. Indeed, big tech itself agrees that there must be some form of regulation for the internet, as it continues to increase its presence in our lives. Even in our divided world, on this, there is consensus.

Sadly for all concerned, that is just about as far as the consensus goes. There were various so-called ‘easy wins’ when it comes to online spaces, but all of these were reached years ago.

Companies should have liability for customer data being lost or stolen due to negligence – sorted. Tech should make significant efforts to keep child abuse imagery off their services – a perpetual struggle, but one in which tens of thousands of images a day are removed. Wherever they are based, tech should remove speech that does not comply with local law – also long ago, sorted.

This is most notable when looking for compliance with Germany’s understandably strict laws against speech denying the Holocaust. Social networks now take down content that falls foul of this law, but is legal speech in other countries, thereby restricting it solely in Germany.

Big tech has made it clear they will comply with democratic laws on speech. They say if you want particular content not to exist, simply make it illegal – otherwise, leave the limits of legal speech up to us as individual services.

Some services, like Facebook, will not let so much as an exposed female nipple be on show – even if breastfeeding. Others, such as Twitter, are happy to host hardcore pornography. Each makes a choice according to what its users and advertisers are willing to accept, and acts accordingly.

But the UK’s approach in the Online Safety Bill would have taken that much further. The Bill created a new category of ‘legal but harmful’ content which social networks would need to demonstrate they had plans to tackle without outright banning it. With absolutely swingeing fines at stake, such a policy would clearly result in most networks zealously over-censoring content, given the costs of falling foul of the law are so much greater than the benefits of supporting an open internet.

Such was the backlash from the tech sector that the Government tried to reframe the ‘legal but harmful’ restrictions as content likely to be accessed by children – on the face of it, a much more reasonable middle ground.

This was essential, as the Home Office had used children’s charities to front much of the legislation and convinced them it was necessary to keep children safe online. In classic Home Office fashion, it set a trap and then immediately fell in it: pass the measure, and big tech withdraws jobs and activity from the UK; do not pass it and face a huge public backlash from charities that are largely above reproach.

The compromise does not work, of course. Almost every service online is likely to be accessed by children, generally defined as a child under 13. Kids can easily tick a box saying that they are older than they actually are and generally are excellent at finding things that they should not.

As a result, the only way to really exclude children is to introduce stringent age verification checks, which rely on some form of real-world ID, akin to those on gambling websites. That has put the government in direct conflict with Wikipedia, the worthy not-for-profit encyclopaedia cribbed by children across the world for their homework assignments. Wikipedia, not willing to introduce an age filter and not able to comprehensively moderate its site, has said that, if these measures pass, it will withdraw from the UK.

If any of this was easy, then it would have been done and dusted long ago. But there is a fundamental problem at the core of the rules: they do not bring big tech to heel. Instead, they merely introduce new restrictions on the ability of private citizens to speak out online.

It is approaching ridiculous if Facebook can be punished for allowing on its site comments that someone could otherwise make in a public space without sanction. The result is not a restriction on Facebook, but rather on the person making the comments.

In general, big tech does not adequately moderate its content, nor does it contribute enough in tax to outweigh the societal harms of its products. The huge profit margins of the tech giants,  especially in their online ad divisions, is telling of a surplus that is causing issues.

Good legislation might look at trying to levy them in proportion to the harmful content on their platforms, with a sound legal definition of such content, or else to require minimum moderation ratios and support times for large sites. But it should not try to outsource the limits of speech to cautious tech compliance companies, which is what the current Bill proposes.

When we do not like what we see online, it is rarely big tech’s fault. We must ensure that when we are trying to regulate business, we do not instead merely regulate each other.

James Ball is a British journalist and author. He has worked for The Grocer, The Guardian, WikiLeaks, BuzzFeed, The New European and The Washington Post. 

This article was published in the latest edition of Centre Write. Views expressed in this article are those of the author, and not necessarily those of Bright Blue. 

Read more from our August 2023 Centre Write magazine, ‘Back to business?’ here.