Just before the COVID-19 pandemic struck Europe, Brussels was en route to designing its most ambitious tech agenda to date. A digital strategy, including the long-awaited Digital Services Act and an overhaul of competition rules, a Data Strategy, covering business to business and business to government data sharing and an AI White Paper, looking into the necessary regulatory framework for developing safe AI in Europe were all launched on 19 February. Fast-forward 4 months and the inevitability of re-prioritisation is happening, with consultations and draft legislative documents postponed. So what stands of the initial plan and what needs to be re-thought in view of the profound changes to our economy and society that the coronavirus brought over? How does a future-proof EU tech agenda look like that rebuilds and reinvents the European economy, while ensuring high levels of protection and security? What does the new contract between people, tech and regulation look like?
This was the topic of the tenth of a techUK series of webinars discussing the world in the aftermath of the COVID-19 pandemic, featuring an expert panel:
Werner Stengg, Cabinet Member, Executive Vice-President Margrethe Vestager, European Commission
Sarah Cameron, Legal Director, Pinsent Masons LLP
Elizabeth Crossick, Head of EU Government Affairs, RELX
Eline Chivot, Senior Policy Analyst, Center for Data Innovation (Moderator)
Eline Chivot started by setting the scene, showing how COVID-19 has re-shaped our work and lives and how the EU tech agenda will need to respond to these new developments. She reminded viewers of the EU Digital Strategy, the AI White Paper and the upcoming Digital Services Act, the initiatives at the core of this European Commission’s tech plans. With some inevitable delays to the legislative process, she noted that all of these initiatives remain very much a priority for the Commission. She then went on to ask what sort of impact the COVID-19 pandemic may have had or should have had on the EU tech priorities and the strategic thinking behind them.
Werner Stengg started his intervention acknowledging that this crisis has shown what tech can do when nothing else works and that it was a crash course for all of us in all things digital. The COVID-19 pandemic has not changed the Commission’s policies, it just showed how ever more important the EU tech strategy is, giving the increasing role of tech in our lives. Along with the green transition, the digital transition is one of the major priorities of the Commission, yet a lot of the green transition is coming through digital solutions. That is exactly why, he said, we need to shape Europe’s digital future, and make sure that all these wonderful new innovative technologies are deployed in a way that serves our society, our economy and our people. The Commission wants Europe to play a stronger role in digital than it may have played in the first wave of digitisation but in a way that reflects European rights and values. He envisions a Digital Market based on values that gives everyone a chance to perform, that is transparent and where information can flow freely. Given how many business are now operating online, due to the COVID-19 pandemic as well, he believes it’s more important than ever to create fair market conditions for small and medium sized players. The Digital Services Act is also meant to make sure that European citizens are not exposed to unsafe products or illegal content. He closed by saying that cyber security and the skills agenda are also top of the Commission’s agenda.
Sarah Cameron reflected on the indebtedness to the technology sector for keeping us going through the crisis, but also for all the tech4good stories that we’ve seen through the pandemic. She stressed that connectivity needs to be seen as an essential service for all so that nobody is left behind. Looking at the recovery phase, it’s more important than ever that investment into low-carbon economy and green tech, digital health, upgrading skills is turbocharged so that governments send a strong signal that money flows to the right places according to our values. However, Sarah noted, there is a mixed picture on how the tech sector is faring. A recent survey suggested that about 60% in the sector thought they would suffer no more than a moderate impact but the small tech businesses are likely to be disproportionately affected; so it's very important that they are helped to survive and thrive and not disappear or be bought up. She emphasized that the EU digital strategy chimes with a lot of the key objectives that the UK is also looking at, around trust in the data economy, digital skills or access to data. On the overhauling of the e-commerce directive, Sarah acknowledged that there is a need to remedy fragmentations in the market and have a discussion around illegal content online and the allocation of responsibilities and liability. However, on the wider competition elements in the Digital Services Act and the parallel competition tool, she thinks the debate is likely to be more intense. On the Data Strategy, she agrees with many of the Commission’s high-level aims and ambitions. She also stressed the critical need to work really closely with industry to make sure that innovation is not stifled by some of those initiatives and regulation is pursued where really necessary, for example where there are public safety risks. A detailed conversation about access to data is required. Data is going to increase 500 percent in the next five years - a potential of EUR 829 billion worth of value to the EU. While she notes that great progress has been made with the public sector, when it comes to the private sector she thinks that rather than jumping to mandating data sharing, we should look to incentivise and assist adopting some of these initiatives, by engaging stakeholders and considering all the relevant commercial interests. Lastly, she said that she’s a great believer in the fact that AI should be looked at as a great force for good, but with risks that need managing and not the other way around. It can create thousands of jobs and drive significant growth, and there is consensus too that most of the existing regulations and laws largely cover what's required for AI, including liability. There may be some gaps and high risks to address, but this will come down to a balance and evaluation of risk. When dealing with attribution and liability between business she thinks there'll be a strong argument for businesses to address this risk between themselves and any conversation around reversing burdens of proof needs a very close consideration with business so that you don't end up stifling the very achievements that you're setting out to achieve. Strong governance frameworks can play a major role that fit with the wider principles broadly agreed around the ethics of AI. The high risks need to be looked at from a sectoral perspective. She concluded by saying the EU work is largely very well-intentioned and if we can have strong collaboration with industry bodies but also internationally on areas of international focus then great progress can be made and we can look forward to a thriving economy.
Elizabeth Crossick started by saying that Europe’s work on setting rules and standards, on building up its digital infrastructure and removing barriers to the Single Market for data should not interfere with openness and data flows. A universal definition of AI would be extremely helpful and there's a useful exercise to be run by the Commission to help bring together resources to improve citizens understanding of AI. It’s really important that individuals feel empowered to interact not just because they have regulatory safeguards but because they understand the benefits not just the risks. That’s crucial for not just a broader uptake but for broader trust in AI. On data, she noted that not all data is equal and the outcomes from AI applications or systems can only be as good as the underlying data that are fed into the algorithms. She added that a robust and flexible IP regime in the EU will help guarantee a successful AI sector by incentivizing creativity and ensuring those breaking the rules aren't rewarded. She highlighted that one of the most important components to help the EU’s acceleration on data is access to public data for which the full implementation of the Open Data Directive by all Member States is imperative. Her second point related to the relationship between AI and GDPR. The White Paper presents various policy options in terms of regulation and there is an extensive body of existing legislation that's applicable but may well need updating. One area she would welcome some adjustment on is the application of GDPR on AI. The white paper identifies specific concerns relating to bias and discrimination but GDPR makes it extremely difficult to test for bias within training data due to the restriction on collecting that sensitive data in the first place. On the DSA, she said that “notice and take down” should me “notice and stay down”, following both European Court jurisprudence and the Copyright Directive which note that platforms must act expeditiously to disable access and make best efforts to prevent their future uploads. She suggested adding that in the DSA would be a logical extension. On liability, she notes there's been much discussion about this distinction between active and passive and when the latter lose their exception. The DSA roadmap refers to this on option 3. She welcomes clarification but cautions against any introduction of the so-called Good Samaritan approach, not least because it hasn't brought any legal certainty in the US where it's best known and she thinks it would totally redefine the current EU legal framework that relies on the e-commerce directive articles 12 to 15 and could leave rights-holders with an overly broad principle that would be difficult to enforce. There's also been some discussion around duty of care: the larger the platform the greater the obligation. It's not clear how one can reconcile this concept without completely changing the liability rules. Furthermore, it isn't always size that matters. A small platform that intermediates the sale of products some of which are unsafe and potentially lethal to consumers should surely not be less responsible than its larger competitor. She closed by saying that any legislation introduced should be technology neutral and avoid being over prescriptive. With the right conditions and through the twin ecosystems of excellence and trust, technology and AI will flourish in Europe.
Cameron Kerry started by saying that one of the drivers of EU policy is a sense that the EU needs to play catch-up, that it's been a laggard in tech developments and the digital economy. However, he did mention and expressed admiration for the EU’s position of leadership in terms of digital policies. He noted that high-level discussions are taking place on the opportunities for transatlantic cooperation on AI between the Commission’s DG CONNECT, White House officials, UK’s DCMS, the Canadian Ministry of Innovation, Science and Economic Development. In spite of some of the tensions and in spite of some of the rhetoric on the US side, he thinks there is a recognition that this is an important area for cooperation. Because of shared values and because the EU and the United States make up about half of the world's GDP and third of the world's trade flows, he believes we are inexorably linked and this is an important opportunity to collaborate. The AI white paper recognizes comparative advantages on what the EU does well in robotics and in machine-based AI, but there are advantages on the other side as well. Canada for example has been a leader in this area, as it's the first country to adopt a national AI strategy. He noted how important the role of the private sector and the role of experts is going to be in all these conversations on regulation, given that government officials are hardly ever technologists. When it comes to the opportunity of scaling in the data economy, Cameron mentioned the case of China and its ability to collect and analyse data in ways that liberal democracies don’t operate. So he thinks liberal democracies need to work together to allow for scaling of the data economy within the same rules and around the same values. He concluded by saying he sees a great opportunity for international cooperation on AI, for harmonisation and inter-operability, because from a regulatory standpoint this is very much a green field.
You can watch the entire webinar below: