MMC interviews Jonathan Ledgard
“AI’s rewards and risks”

The following interview was originally compiled for the Mixed Migration Review 2019 and has been reproduced here for wider access through this website’s readership.

We may be several decades off from extremely advanced and all-pervasive multi-tasking forms of AI, but narrower types of the technology are set to spread exponentially over the next few years, says Jonathan Ledgard. That’s good news for pretty much everyone in countries both rich and poor, except migrants and refugees.

Where are we with AI and are we close to General AI or even ‘superintelligence’?

General artificial intelligence? Not in our lifetime, not in any meaningful way. But what you will see within the next five to 10 years is pervasive narrow artificial intelligence. So, thousands of narrow AIs, and we already have many of them, but within the next decade they will be very pervasive. It could be energy efficiency or cancer detection or meteorology prediction, and indeed then you end up with these merging into the sensory systems, both natural and tech-based. So, for example, smartphones become empowered, and have already started to become empowered by these narrow AIs, and the way that they see and the way that they listen becomes very different.

In the context of a migration discussion, we have to accept that facial recognition will be pervasive, even on smartphones. So the poor and the vulnerable will be tracked. Even if we don’t know their names, even if we don’t know where they came from, or we’re not quite sure where they’re going, they will have an identity, and they will be plugged into the system. AI will give us the possibility of curing diseases and the possibilities of optimising almost everything, but then also the possibilities of a police state too.

In the AI technology race, will the winner take all and cause significant disadvantage to developing economies which may not have an initial stake? Will it lead to further global inequality?

I’m more optimistic. My personal view is that the lower down you are on the spectrum, the more positive the influence of both artificial intelligence and robotics can be in your life, because you do not have access and you’re not able to afford access to how markets function normally. You don’t have the ability to go to a bookshop or a library and so on and so forth, you don’t have the ability to have pretty high-end healthcare. Then I think you will see that artificial intelligence can be a significant improver in quality of life in many ways, in sub-Saharan or South Asian environments.

For example, we could talk about intelligent drip-feed irrigation, which can turn a hosepipe into something quite sophisticated using simple AI or a smartphone, or we could talk about recognition of disease, early warning of cancers. Now, of course, there’s no real cancer care in most African countries, but I think it’s going to improve, dental care will improve, eye care will improve. In terms of the use of energy, agriculture, healthcare, also, optimisation of transport in cities, I think artificial intelligence can really be a point of play.

It’s a complex and subtle subject to unpack completely in a short time, but on balance, I think it’s going to be a net gain for poorer countries, not a net loss.

What is your vision for and of drones in the future?

I see drones only as the first accessible scalable robots. I’m interested in drones as a robotic platform in which forms of AI can empower them. We have drone legislation in six African countries. And now in Rwanda, over half the blood in the country is flown by drones. You will see that 3%, 5% of high-value goods will go in the sky and that, especially for secondary towns, can be quite a useful addition for them and for the healthcare sector, especially, we’ll probably see some improvements in the supply chain. We are now exploring with the World Bank and others whether we can actually start building droneports in the next two or three years.

What exactly is a droneport, and how do you see its role evolving?

The key rationale for a droneport is that we don’t think it’s credible or safe in any way to have drones just flying willy-nilly without any oversight from the government. Militaries, intelligence services and the general public won’t accept that. And therefore, the cargo drones, which will be relatively large, will be carrying small suitcase-size loads through the sky. They should be known to the government, and therefore they need a safe place to land and take off. That will be the droneport, situated right in communities, and part of community life. And that’s the first stage. In order to have a scalable system, that system has to operate within some parameters. And that probably means you’re going to have fixed routes in the sky.

They are not going to replace all other forms of transportation, but they just add something very cheaply and tangibly. The retail sector, the shopping sector in Africa will not be built out the way it has been built out in Europe or in the US or in Japan. There will be some high street shops, but people who want a choice of valuable goods are likely to access that choice on the internet. And because they access it on the internet, the supply chain can optimise for drone delivery, but not last-mile. Last-mile should be human, motorbike and so on. This is how droneports are currently envisaged. There’s a regulatory hurdle, but the main hurdle is the battery technology.

Can you see any benefits for migrants and refugees from AI and drones? Or are they most likely going to be used to keep them out?

On this I’m much more pessimistic, unfortunately. In general, I think, for anyone who is genuinely displaced or a refugee, they’re going to be deeply damaged by this technology.

As much as we can imagine a lot of positive upsides, in a civilian environment where people are on the move from where they’re very vulnerable… I think politics around migration will push the militarisation and securitisation deeper into the Sahara, for example, through every bottleneck that there possibly is, and it would be very cheap and easy for governments to do that.

There is a paradoxical element here insofar that you may not know who a particular woman on the move is, and you may not care in any way who she is, but she will be in your database very quickly, and AI is not helpful in that sense.

How do you see AI affecting demand for labour with regard to migrants and refugees?

On this, I’m a little bit more optimistic. I do think that there will be greatly increased demand for more skilled agricultural labour, and more skilled healthcare sector labour. Those are the things which AI and robotics won’t get into in a meaningful way. Robots are already pushing hospital beds around in Japan, but they won’t be able to lift a patient, and change the sheets, and comfort people, and so on. I can really imagine on the semi-skilled to skilled side of both the agriculture and caring industries there will continue to be quite a lot of opportunities. Where it obviously gets difficult to predict and very multi-dimensional is the issue of job displacement of middle-class jobs in wealthier industrial countries. It’s going to be very combustible, and it’s hard to play out all the scenarios.

What about the longer term implications of AI?

Most of my work on AI is to do with AI and perception of nature, or how AI perceives other life forms. What does artificial intelligence look like if it has no imagination about the living world at all, or things which are living in this world which have no utility to the human economy? And I think it’s deeply alarming and worrying that artificial intelligence in the moment is evolving, and the architects of AI are evolving it, without real context of being on a living planet.

Now, having said that, I definitely believe that, say by 2060, 2070, that Homo sapiens as a species will
have meshed, to some degree, with forms of artificial intelligence. Whether they subsume those forms to increase longevity, intelligence and sensory perception or whether they merge into a new species, which is probably what I believe. I’d be really surprised if Homo sapiens is around in two centuries’ time. I would be really surprised if we exist in our present form. It’s just simple mathematics. The AI programs are operating 16,000 times faster than the human brain can operate. And as much as the human brain has this plasticity and this incredible breadth of mobility, that’s not going to be enough. And so I think we’re definitely living in really profound, historic times.

How do you view the future? Are you pessimistic, optimistic, dystopian or utopian?

The thing I learnt from living in Africa was that pessimism is a waste of time. It’s a waste of energy and a waste of time. I’m optimistic, I probably should be more realistic. Concerning the changing nature of Homo sapiens, I’m not particularly alarmed by that. I think there are many reasons to think this might be a more benign future as well.