Hello and welcome back to the Pinsent Masons podcast with me, Matthew Magee. We hope to bring you the latest business law news and analysis from our experts around the world every fortnight, and this time we’ll hear about the funding structure being used to funnel billions into sustainable development, and we look at some of the trends emerging in the earliest court cases about how AI actually functions.
But first, some business law news from our Out-Law reporting team.
Hong Kong budget emphasises safe, green construction
Some Apple and Microsoft services escape more intense EU regulation, and
Qatar court upholds contract’s mediation clause
Hong Kong Special Administrative Region (SAR) will increase infrastructure spending by 17%, its administration has said. It will focus on the Northern Metropolis development, a project intended to be an international innovation and technology hub.
One expert said that the infrastructure spending priorities in the region's 2024-2025 budget will direct spending towards projects which incorporate technology, prioritise safety and create green infrastructure.
Hong Kong-based infrastructure expert Alvin Ho said that to maintain or win more market share, infrastructure companies must embrace greater technology application, improve site safety and reduce environmental impact in the way they operate.
Some services operated by Apple and Microsoft will escape more stringent regulation in Europe after the European Commission ruled they are not 'gatekeeper core platform services' under the Digital Markets Act.
This means that the four services will not be subject to additional competition rules and compliance obligations aimed at the EU activities of gatekeeper digital platforms.
Apple's iMessage and Microsoft's Bing, Edge, and Microsoft Advertising services have escaped the increased regulation, but other services offered by Apple and Microsoft, including Apple's app store and operating system, and Microsoft's LinkedIn and Windows operating system, will retain their earlier designation as gatekeeper core platform services.
Qatar's Investment and Trade Court has ruled that companies could not go to court until they had conducted mediation as laid out in their contract in a ruling which one expert said could change how companies there handle disputes.
The court said that litigation could not proceed until mediation had happened because the contract said that disputes should be settled by mediation. It considered the issue under Qatar's mediation law and said that a mediation clause should be treated as the same as an arbitration clause in a contract, and that the clause meant that it did not have authority to act in disputes where mediation clauses exist.
Arbitration expert Pamela McDonald said that companies operating under contracts with a mediation clause should now assume that unless they have attempted mediation of disputes, they cannot commence litigation proceedings.
When the nations of the world gather to discuss climate change and what can be done, one of the things they talk about most is how to pay for all the changes that lie ahead. Building factories to make wind turbines or solar panels; retrofitting buildings; creating new energy or water networks for a changing climate and changing how we produce our food all have one thing in common – they need money, lots and lots and lots of money.
There is grant or philanthropic or governmental money out there, but not enough by a long chalk. Enter blended finance, a way of structuring investment funds that seeks to use that relatively small amount of money to incentivise more profit-focused investors to invest in riskier but climate-beneficial projects.
Glasgow based investment expert Elaine MacGregor told me how blended finance works.
Elaine MacGregor: It's a way in which you can get investors with different objectives to invest alongside each other. So, you have your impact focused investors investing alongside investors who may have more of a focus on financial returns and making sure that they can be involved in the same project or the same fund, but there have been some tweaks and adjustments to how they're involved to make sure that they're all meeting their own objectives as well.
Matthew: Ok, so a more traditional fund is where everybody just puts money in and they get returns proportionate to the money that they put in or loses proportionate to money they put in. So how is blended finance different from that?
Elaine: Normally, investors will come into a fund on the same basis and achieve returns on the same basis, but in a fund you can actually separate out different types of investors, so you may have those who are more comfortable with risk or those who seek different returns and in that way you can use blended finance to adjust what different investors want to achieve and make sure that you're bringing in the commercial monies that would be needed to help a fund achieve its sustainable development goals. What a blended finance investor tends to do is look at what's stopping a commercially focused investor from coming and becoming involved in the fund or the project. It may be that the commercial investor thinks it's far too high risk, so the blended finance investor will come in and say, you know, we're happy to take the first hit on losses. So, we may structure the fund so that they could be paid later or there could be some grant change that takes the losses first, the blended finance investor, they can come in and say look, we're happy to target lower returns for ourselves so that you can achieve the higher returns you want if you want to get involved here. It's all about leveraging and mobilising commercial monies, because just rely these, you know, these are finite resources we have from our public and foundational investors and you, you need a lot more money to coming in. So, it's trying to be smart about how you use your money and making sure that you're catalysing or mobilising more commercial monies to get involved as well.
Matthew: So, I think we can probably imagine who the commercial investors here are. It's the people who've been investing things for decades. Who are these impact investors that are happy to take lower returns or to make losses first? Where's that money coming from?
Elaine: So, you have different types. So, the general categories you have public investors. Now that could be sort of International Development agencies USA does a good example. Other public investors could be government development, finance investors, so UK's equivalent is British international investment. You also have philanthropic investors, more charitable and the most well-known example, I think, is the Bill Gates Foundation. They're heavily involved in a lot of sustainable development goals.
Matthew: The United Nations in 2015 created sustainable development goals, 17 targets that, if met by 2030, would create “peace and prosperity for people and the planet, now and into the future”.
Developing countries in particular need funding to undertake the projects that would help them meet these goals. Blended finance can help.
But it is relatively new and still developing. When you hear that $200 billion is invested in this way it sounds like a lot, but in the context of the amount of money needed to meet sustainable finance goals it is still pretty small.
Elaine: Convergence, which is a Canadian blended finance initiative, they've estimated that there is a funding gap between what developing nations need to achieve climate related actions as well as sustainable development goals, and for this year it's estimated at $4 trillion, that's a big gap to overcome. It's a hurdle to achieving those sustainable development goals and certainly I think to date it's been estimated by convergence that over $200 billion has been leveraged through the use of blended finance. So, it's certainly a start to meeting that funding gap.
Matthew: Blended finance isn’t all about cash money. With many of the world’s biggest development organisations and philanthropic funders involved, there is the opportunity to get projects up and running with all sorts of different kinds of support.
Elaine: But it can take the form of technical assistance, so some of the impact investors can provide training, for example on ESG standards, or they can also help on operational matters like framework reporting and also they can provide maybe some grants to assist with initial feasibility studies just to get things kick started. Another way blended finance can help is actually by reducing financing costs. So blended finance investor maybe could provide guarantees or insurance on very generous terms. Certainly, we see more innovative solutions. One of the latest we've seen is around selling fund interests that have been held for a few years, so that's using time to reduce risk. So in a way, the impact investor took the initial risk of making the investment, but a commercial investor can step in and acquire those interests. They will achieve distributions more, you know, much more quickly than they otherwise would have.
Matthew: And are there downsides to this kind of investing? Are there risks? Maybe for the projects themselves or the countries where they're located?
Elaine: Yeah, there are risks. It's a delicate balancing act I think for any blended finance investor. They obviously want to avoid having excess subsidies, so if you're providing monies at very generous rates, you could potentially be crowding out commercial monies. But also on the other side, you want to make sure that you are providing enough, so it's sufficient to mobilise the private monies, because at the end of the day, you still want to create that development impact. So, it is a delicate balancing act and I think every blended finance investor has to consider those aspects carefully when they make their structuring arrangements.
When Steven Moffatt suffered a bereavement and had to book a flight he was careful to check the information about Air Canada’s discounted bereavement fares. The Air Canada website’s AI-powered chatbot told him he could claim a refund later and pointed him to a page with more info. That page said he couldn’t claim afterwards. When the airline refused to refund part of the fare later he sued, and won.
Air Canada said that it could not be held liable for information provided by one of its agents, servants or representatives, including a chatbot, but Canada’s Civil Resolution Tribunal said it was ‘remarkable’ that Air Canada would claim that its chatbot was responsible for its own actions and that the company was liable for all the information it provided, whether via a chatbot or its website.
Now this is a pretty small-stakes case in a pretty low-level court, but it neatly illustrates the anxiety of people, courts and companies: when using AI systems the whole point of which is that they make their own decisions, who is responsible when those decisions are wrong? Is it the company that used the system? The company that invented it? Or the person who relied on it without double checking its results?
London-based technology disputes expert Meghan Higgins followed the case.
Meghan Higgins: The court found that the Air Canada chatbot had made a negligent misrepresentation about the Air Canada bereavement fare refund policy and the claimant got his money back from Air Canada. Air Canada said that it was not liable for it, the decisions of its representatives, and essentially that the correct information had been available. They didn't really explain their reasoning, but I think there's sort of two points that they're making, one of which is kind of analogizing the chat bot to an actual employee of Air Canada and sort of suggesting that also that when a customer is interacting with the chat bot, it might be sort of their responsibility to go and verify the information provided by that chat bot elsewhere on the website which the court did not accept.
Matthew: What AI litigation has been so far has mostly been about copyright and the data that's been used to train the system. The Air Canada's case has got a lot of attention simply because there isn't that much litigation yet about the outputs of AI and what litigation there has been has tended to focus on how AI interferes with long established personal and informational rights.
Meghan: One of the areas that we're seeing these disputes is in automated decision making, which under Article 22 of the GPR, there's various constraints imposed when you deploy a system using automated decision making that can have an impact on people's legal rights. There has been a decision this year from the Court of Justice of the EU about a credit agency that generated credit score using data fed into it and data subject challenge that because that credit score was used by a lender to assess whether or not the data subject would get a loan. Automated decisions are becoming so pervasive in our day-to-day life and we don't always have visibility as to when these systems are being used or how they operate and you know in a lot of cases, companies relying on the systems may not have visibility into how how they operate really or how they reach a particular decision or output. So, I think we will see scrutiny of those kinds of systems going forward and people challenging the outputs of those systems, if they impact their rights. There have been some human rights claims about the deployment of AI systems, particularly against the public sector so where, for example, algorithms or automated systems are used to allocate benefits or to test eligibility for particular programmes etc. These tools are being deployed, but it's not always possible to understand kind of the risks associated with deploying these tools, particularly in sort of a context such as distribution of government benefits that has a really high risk to peoples human rights.
Matthew: What this Air Canada case really highlights is the conundrum facing courts. Is AI a person, like a worker, or a tool, like a piece of software or a forklift truck? Liability in those cases is really clear and is based on the foreseeability of mistakes or damage. But that doesn’t – by definition can’t – work with AI systems, whose essence is that they create and generate and develop, so the implications of their use is unforeseeable.
Meghan: Typically somebody in the position of Air Canada or a contracting party is only going to be liable for a loss that would be reasonably foreseeable to the party and so I think this is where you get these really interesting questions with AI, where the output of an AI is not necessarily predictable or foreseeable and how we sort of fit that kind of tool into these more traditional legal concepts. Those systems that are fairly predictable are probably more analogous to a traditional piece of software or a traditional tool. If, for example, Air Canada had licenced this chat bot from a third party, it's possible that Air Canada did not really know what some of the risks or issues associated with this chat bot might be. Just because the technology is really complex.
Matthew: And actually we know because some of the inventors and founders have said this, that even the people making the AI systems find it very hard to predict how they behave or how they work in the future, which is presumably going to present a really, really difficult question for courts. We're used to at least the people who come up with systems knowing how they're going to work.
Meghan: You know, we have these decisions that go back where somebody commits a tort and there's a completely unforeseeable consequence that, you know, a chain of events that nobody could have ever predicted and the courts have really drawn a line there and said, well, we can't impose liability there because essentially, that would make either the person who committed the tort or had breached the contract, the insurer of any possible loss that might arise from that legal relationship, and nobody would take that on nobody, it would be much more difficult to kind of agree a contract if that happens, because that's just there's too much risk there.
Matthew: So this leaves everyone, people, companies, courts in a bit of a bind. How will they navigate the complexities ahead? Perhaps through a mixture of judicial fudging and new legislation, thinks Meghan.
Meghan: I think courts will adopt an approach where they are trying to sensibly allocate risks to the party who can best protect against them but courts may have to adopt some either some presumptions or we might see some sort of judicial developments that kind of give us more of a steer about how these cases will be dealt with just as they arise and courts kind of look at these issues and realise that the existing paradigm we have as to foreseeability doesn't really work on this context. We could also see sort of legislation emerging that kind of seeks to kind of allocate certain responsibility, and I think we have had that in the kind of driverless car situation in the UK and probably in other jurisdictions as well.
Matthew: So lots of companies are using AI in all sorts of ways, and they're trialling it some with their own processes, some with their own customers. They might be a bit alarmed by this ruling in Canada. What can they do now to prevent trouble in the future?
Meghan: Just thinking about the kind of arrangements that companies are entering into to procure these tools and the kind of contracts that underpin that procurement and how risk is allocated in the contract, the kind of liability that could arise, which is really difficult because we are in the early days of these cases. But thinking about that and making sure that there are provisions addressing that in the contract as well.
Matthew: Thank you for listening again. We know that you have a lot on. We know there's a lot of information out there so we really appreciate the time you spend with us. If you think this podcast would be interesting or useful to your colleagues contacts or friends, then please do share it. And remember, you can get up-to-the-minute business long use and analysis from our team of journalists at pinsentmasons.com and make sure that you never miss any of our news by signing up for updates at www.pincentmasons.com/newsletter. Thanks for listening and see you next time.
The Pinsent Masons podcast was produced and presented by Matthew Magee for international professional services firm, Pinsent Masons.
