This piece looks ahead to what we might expect as IT law developments in 2019.
You can tell that digital transformation really is well under way when commentators who in past years have freely given just 10 top IT predictions for next year start to signal ’31 Tech Predictions for 2019’ (inc.com), ’60 cybersecurity predictions for 2019’ (Forbes) or put their entire set of 23 2019 prediction reports behind a paywall (Forrester). So, at the risk of over simplification, I’ll nail colours to the mast and go for the cloud, AI and data as what’s coming up for IT lawyers in 2019.
First, though, let’s clear the Brexit decks. At the time of writing (early December) withdrawal from the EU at 23.00 on 29 March 2019 is planned to be followed by a transition period ending on 31 December 2020 so, other things being equal, the UK’s legal relationship with the EU during 2019 (and 2020) will be largely unchanged from today. If however Brexit takes place without agreement, then IT and IP lawyers will be grappling after 29 March with intricate technical legal questions of divergence particularly around GDPR, copyright, database right and exhaustion of rights (the EU law free movement of goods principle that IP right holders cannot assert their rights to prevent resale of their goods once sold in the EU with their consent).
Quantum Computing (QC) and Blockchain
Further out, QC is set to change the fundamentals of computer processing (and Rigetti Computing launched first QC as a Service (QCaaS) in September 2018), but it will be after 2019 that QC uptake starts in earnest, with cryptography and security likely the first commercial use case areas. Similarly, although blockchain has the capacity to reshape high volume, low value record-based business operations (like asset tracking, identity management, patient health records and B2C insurance) and there are already around 100 commercial applications for the technology today, it will be after 2019 when blockchain enters the commercial mainstream.
The first of two IT law predictions I’d be pretty confident about for 2019 it’s that there will be many more cloud contracts coming across IT lawyers’ desks next year. A central feature of digital transformational is the epic migration now underway in enterprise (large organisation) computing from ‘on premise’ – traditional IT infrastructure at the user – to ‘in-cloud’. Wikibon, the open source technology knowledge sharing community, projects that the cloud’s share of enterprise computing will grow from 10% currently to 45% by 2026 (see Figure 1 below).
An enterprise cloud model emerging strongly at the moment is the hybrid cloud. This harnesses the scale of large, ‘core’ data centres with smaller ‘edge’ data centres (off-premises, but located closer to the customer where the applications and data are used). Enterprise customers can choose from the range of Infrastructure (IaaS), Platform (PaaS) and Software (SaaS) cloud services but are increasingly (especially in regulated sectors where security is critical) installing and managing their own infrastructure for their hybrid cloud operations. Here, vendors (like Dell/EMC/VMWare, now effectively under one corporate roof) are increasingly integrating their server, storage, networking and software offerings as ‘Hyperconverged Infrastructure’ (HCI) to deliver easier configuration and integration, greater flexibility and reliability, and lower costs.
For IT lawyers, there is a range of contracts to be put in place when an organisation first goes to the hybrid cloud, especially when it’s putting in its own HCI rather than buying IaaS. Effective contract structuring and negotiation play an important part in the success of the project, where the key questions include:
- what is the world of contracts to be signed up (agreements with data centre, telecoms, equipment, software and services vendors)?
- what are the dependencies between them?
- what legacy contracts are to be terminated?
- who does the customer look to if something goes wrong during integration or after go-live? Does it have ‘one throat to choke’ or is it doing its own management and identifying and then looking to a particular provider to fix the problem?
- what’s the exit plan?
If getting the contracts right is one half of the cloud puzzle in 2019, the other half for IT lawyers is putting in place the right framework and governance for cloud security and compliance to manage applicable cloud security duties. Here, enterprises are looking at consistent approaches to best practice, assessing cloud risk and obtaining satisfactory vendor assurance.
Figure 1: Worldwide Enterprise IT Projections by Segment, 2017-2026 ($bn). (Source: Wikibon)
AI – a set of technologies not a single one (see Figure 2) – will move particularly fast in 2019. In the last few days alone, AI developer DeepMind has announced how its unsupervised machine learning AlphaGo Zero program, starting with only the rules of chess, thrashed the top chess program; and, in the area of natural language processing, the editor of the FT has said:
“[w]e are rapidly approaching the moment where all text can be understood by machines — a revolution as big as the launch of the internet.”
Research consultancy Gartner recently predicted that business value derived from AI will grow annually at 50% over the next few years (from $1.2tn in 2018 to $3.9tn by 2022). As with the cloud, this means we’ll be seeing more AI-related contracts next year.
Figure 2: The main AI streams
Emerging business models include the AI machine learning platforms (AIaaS) of the large cloud providers (Amazon, Microsoft, Google and IBM Watson) where pricing is determined by number of compute hours, transactions and training units.
The customers of these AIaaS platforms are large end user public and private sector organisations doing their own AI development, and the thousands of specialist AI providers Gartner forecasts will proliferate between 2019 and 2022. The business model for these niche AI providers and their end user customers centres on (cloud) SaaS subscription or (on premises) software licensing (for the AI software) and professional services (for dataset testing, training and support).
Where the end user customer provides its own datasets (a law firm using its ‘secret sauce’ for its AI use case for example) pricing is relatively straightforward. But pricing can be more complex for commercial data vendors (say in healthcare or market data) providing their datasets to a customer for an AI project. This is because, like learning to ride a bike, once the customer’s AI algorithm has been trained on the dataset, that training cannot be ‘unlearnt’ whether or not the vendor’s dataset is used again. Pricing these ‘one shot’ dataset use cases to avoid cannibalisation where the benefit derived is ongoing is currently a pressure point in modelling these types of AI deals.
40% annual data growth means that the volume of global data will double between the start of 2019 and the end of 2020 – and it is this growth that is fueling the Cloud, AI and the fourth Industrial Revolution more generally.
The second prediction is it’s a racing certainty that 2019 will see the first significant fines under GDPR. It will be interesting to see how much they are. The pattern from the competition law world (where the limit under EU law is 10% of worldwide turnover rather than 4% under GDPR for larger organisations) is that fines have risen steadily over time: fines have averaged $2bn a year since 2013, up from $0.8bn between 1990 and 2012. Will the data protection regulators start small or large?
As well as UK and EU regulators getting to grips with GDPR, the more general trend towards closer regulation of the privacy and security of personal and other data will continue to gather pace and remain in the headlines in 2019, and on many fronts: geopolitically, between China, Russia, the EU and US (watch out for 2019 developments in the Privacy Shield tug of war between the EU and US); competitively, between Microsoft, Apple, Amazon, Google and Facebook; sectorally, especially in healthcare and financial services; commercially, in the GDPR-unleashed ‘data processing exhibit’ battle of the forms in day to day business; and internally, inside the organisation.
What does this all mean for data counsel in 2019? As GDPR settles down in 2019 to ‘business as usual’ after the frenetic preparatory activity of 2017 and 2018, most companies have now resolved their approach to GDPR compliance. Internally, the DPIA (Data Protection Impact Assessment) is emerging as the most useful tactical tool to plan, mediate and document the privacy implications of projects and operations. DPIAs will proliferate in 2019. At the strategic level, we are starting to see more comprehensive data governance within the organisation. Businesses are beginning to assemble what they already have – GDPR governance, data security frameworks, AI ethics best practices and data sharing and trust frameworks – into a more holistic, ‘end to end’ approach to data governance, where the organisation moves out from simply responding to legal compliance to a more proactive, ethical approach based on ‘doing the right thing’.
UK online retailing
A parting thought. UK online retail sales first reached £1bn in 2000, getting on for a generation ago now. By 2012, they had grown to £33bn, or 9% of total UK retail sales. Between 2012 and 2017, they grew by 15% a year, doubling to reach £67bn (17% of the total) in 2017. This accounts for a lot of empty shops on the high street today.
15% annual growth implies doubling in 5 years, so at this rate UK online retail sales will be 34% of the total in 2022 (assuming no growth in the overall market). 15% growth again in 2023 will add 5 points to online (taking it to 39%) and taking those 5 points away from the high street, which will reduce from 66% to 61% (or by 7.5%). Continued 15% growth in 2024 would add 6 points to online (for a share of 45%) and take them away from the high street (reducing it from 61% to 55%, or by 10%). Although it takes time to get there and rarely goes in a straight line for long, growth in online at these rates, already clearly visible on today’s high street, will have an increasingly significant impact with each passing year.
Retail sales was one of the first sectors to start the digital transformation journey. In many ways it sets the example for what will happen in other sectors where digital in 2018 is a tiny part of the whole but growing at a prodigious annual rate. These IT-driven changes and their consequences will be with us long after Brexit.
 For further information, see our June 2018 white paper, ‘Legal Aspects of Cloud Computing: Cloud Security’ – http://www.kempitlaw.com/wp-content/uploads/2018/06/Cloud-Security-White-Paper-KITL-v1.0-June-2018.pdf.
 ‘A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play’, David Silver, Demis Hassabis and others, Science, 7 December 2018, Vol. 362, Issue 6419, pp. 1140-1144 – http://science.sciencemag.org/content/362/6419/1140
 ‘Too big to fail: FT editor Lionel Barber on the future of financial journalism’, Speech given for the James Cameron Memorial Lecture, FT, 23 November 2018 https://www.ft.com/content/d2a3e50e-ef07-11e8-89c8-d36339d835c0
 ‘Gartner Says Global Artificial Intelligence Business Value to Reach $1.2 Trillion in 2018’, John-David Lovelock, Research Vice President, Gartner, April 25, 2018 – https://www.gartner.com/newsroom/id/3872933
 FullAI at http://www.fullai.org/short-history-artificial-intelligence/ citing Thomson Reuters as source.
 For further information, see our September 2018 white paper, ‘Legal Aspects of Artificial Intelligence’ – http://www.kempitlaw.com/wp-content/uploads/2018/09/Legal-Aspects-of-AI-Kemp-IT-Law-v2.0-Sep-2018.pdf.