The phrase ‘AI for good’ is often used, but its implications are usually shrouded in technical jargon. For the uninitiated, this term can seem jarring, and the benefits that can be derived from state-of-the-art artificial intelligence (AI) innovations and their applications are not always easily understood.
In the first part of this conversation with Unravel, Scott Hunter, cloud value advisor at Google, speaks with Siddharth Poddar and Shivaji Bagchi about what ‘AI for good’ really is, and its implications on food security, healthcare and supply chains. He also talks about the role of AI in the development of emerging markets and in taking financial services to an underserved population.
Unravel: ‘AI for good’ is a phrase we have been hearing of for a while. What does it mean to you?
Scott Hunter: ‘AI for good’ is a very interesting concept and its applications are manifold. We are starting to see its use in areas like precision agriculture. For instance, the United Nations declared back in 2009 that we have to double our food production by 2050. However, we only have so much land, and water is a finite resource too. So the ‘AI for good’ we are starting to see is one where we can actually look at maps. We can install monitors in the fields and start monitoring variables like ground temperature and moisture content. Already companies like Bayer and Monsanto are undertaking some vertical integration. They’ve even bought their own satellites. They can now help farmers by providing the right seed for their fields or perhaps the right fertiliser mix to improve yields.
In fact, now we are starting to see many companies actually invest in tech to improve the yield and fulfil food requirements around the world.
We also see ‘AI for good’ in food storage, food movement and other technologies like blockchain. It can, for instance, help understand how cold a refrigerated van was or whether anybody opened it while in transit. Basically, it allows data to be collected for the entire journey of a crop – from field to fork.
What is also amazing is that there’s a tremendous amount of food lost because of spoilage. If you can reduce spoilage, you help narrow the gap in food output levels that must be met by 2050. So if you can reduce the spoilage—which currently stands at about 30%—then it does solve some of our problems about feeding societies and populations.
Unravel: Can you provide another example of ‘AI for good’ application?
Mr Hunter: Another great example is from India – a Bangalore hospital that specialises in cardiac surgery. They have about 50 adult cardiologists and 50 paediatric cardiologists. The hospital performs about 17,000 cardiac or thoracic type surgeries every year. To put that into perspective, all of the UK performs about 35,000 such surgeries annually. What they’ve been able to do at Dr Shetty’s Narayana Healthcare is to understand the procedures, the drugs, and then they’ve been able to put all of this into a very large data lake. Now, they can actually perform surgeries for about $1,900 per patient with a mortality rate of about 10% – a figure similar to that seen in the US.
Dr Shetty’s organisation has begun using AI to provide better services at lower costs for an underserved population. It is also providing digital imaging, with images now examined using AI and machine learning (ML) – about 99% of the digital imaging results are actually read by AI and ML. And they have tremendous accuracy. In fact, they only use a radiologist to read the 1% that fall outside this norm. So those are a couple of ‘AI for good’ examples.
Unravel: Since you are a supply chain expert, how is AI a gamechanger when it comes to supply chains?
Mr Hunter: A lot of change has taken place in the last 20 years. So from my personal experience, we actually did implement certain software on the premises of oil and gas and a hi-tech semiconductor company to increase efficiency. The software was great and many companies still run it on their premises, but it typically only dealt with the four walls within the factory, and then maybe some smaller pieces of supply chain.
Now what we’re seeing is those software being uploaded onto the cloud, and it’s being applied in the control tower using AI and ML. What it does is it allows us to see and wring out costs in the supply chain from the supplier’s supplier, all the way to the end customer. We can perform predictive analyses based on weather, temperature and about where they need to steer the product to. So now it is possible to know if you have a shortage and where you can divert the product from.
Now what we're seeing is those software being uploaded onto the cloud, and it's being applied in the control tower using AI and ML. What it does is it allows us to see and wring out costs in the supply chain from the supplier's supplier, all the way to the end customer.
This end to end supply chain visibility and the ability to calculate and recalculate very quickly – I do think this will change how much value and inventory is stuck in the supply chain. And I believe it’s in the billions of dollars. Many companies are getting into this space and it will be an area that’s going to transform rapidly over the next two years or so.
Unravel: What about the role of AI in reaching financial services to the underserved and unbanked?
Mr Hunter: We’re starting to see an array of digital banking opportunities come up. Therefore, instead of establishing a ‘brick and mortar’ bank in rural areas, we’re beginning to see more banking being done on mobile platforms. There are millions throughout ASEAN that don’t have bank accounts, and with mobile banking, they can be more easily serviced.
With AI and ML, you can have a digital bank, a digital wallet and banking can happen on a smartphone. It opens up more opportunities, not only for customers to sell their products, but also for banking customers to use the bank and buy items. This not only fuels the economy but also helps reach an underserved population subset.
AI and ML in banking also lend themselves well to better anti-money laundering (AML) practices. As we start to go through more and more digital transactions, we need to understand that criminals will launder money. And the only way you can catch that is by running AI and ML models across the top of a banking platform.
Within banking and insurance, we are seeing life insurance firms provide real-time quotes – people can enter their weight, sex and the like and they can receive a quote for the premiums they need to pay. The actuarial tables run in the background and these can provide a quote in almost real-time. That is a vast improvement from having to wait a couple of weeks to get a quote back from an insurance company, like we used to do.
We are also seeing insurance companies move towards digital banking that can be done on smartphones – the types of applications that run with AI and ML engines. Insurance companies are also looking at consumers’ lifestyle and health. They are giving away bonuses based on how you make lifestyle choices. And all of this can be done by just connecting to your smart devices, like an Apple watch or a Fitbit. So, insurance companies are changing to take advantage of this new capability with AI and ML. What is worth highlighting is that all these new technologies are much more customer-centric than anything we’ve seen in the past.
Insurance companies are changing to take advantage of this new capability with AI and ML. What is worth highlighting is that all these new technologies are much more customer-centric than anything we’ve seen in the past.
Unravel: How can AI contribute to economic development in emerging markets?
Mr Hunter: In several ways. But let me give you one example. I am working with a company that is engaging with agricultural and livestock farmers. We’re exploring how we can optimise, let’s say, coffee or coconuts. How can we optimise palm oil production or chicken farming? Typically, farmers sell their products to a middleman or a cartel, and these cartels will squeeze the farmer to derive the best price for themselves. Mostly, they don’t have best price for the farmer in their mind.
One way to actually raise the GDP is by creating co-operatives that can look at the prices of commodities such as coffee, palm oil or chicken and are able to sell at or provide a base cost or base price back to the farmer. The farmers, therefore, enjoy selling their products at a higher margin than they otherwise would to middlemen. The other half of the co-operative then optimises the price and they can sell for a profit as well. But that extra profit is again shared with the farmer, back to the underserved who can actually use it more.
This can’t be done unless you run an AI platform across the top of the supply chain and understand the supply-demand match. So I think this is a fabulous way to lift economies and share profits back with the farmers who can be very poor. I think it is an eloquent idea that can actually benefit countries by lifting large portions of their farmers out of poverty.
The second part of the conversation can be read here.