INTERVIEW | Alexiei Dingli: 'Malta ideal place to launch and deploy AI solutions'

Professor Alexiei Dingli is a Professor of Artificial Intelligence (AI) at the Department of AI within the University of Malta. He has been conducting research and working in the field of AI for the past two decades and formed part of the Malta.AI task-force, chairing the working group on education and workforce. He spoke to BusinessToday about his vision for an AI-embracing Malta in the years to come

Alexiei Dingli
Alexiei Dingli

Malta recently launched its National AI Strategy. What is the vision for the country? And does the government have a role in enabling this?

The Vision, as set by the Prime Minister, is to make Malta one of the top 10 countries in the world with regards to Artificial Intelligence advancements. This was the brief that guided us when we were drafting the Malta.AI National Strategy. The various components of the strategy are divided into two; the enablers and the strategic pillars. The enablers are the fundamental components which make up the strategy. Without them, we cannot build AI systems. These include the underlying infrastructure, the legal & ethical framework and the education and workforce enabler.

The strategic pillars rely upon the fundamental components. These include the private sector, the public sector and the investments (which includes start-ups and innovation). Together, the different elements make up the AI launchpad strategy. The general idea is to make Malta the ideal place from where to test and deploy AI solutions worldwide. Government has a big role to play in this strategy because it is providing the necessary boost, which our country needs in order to gain the necessary momentum.

Artificial Intelligence, Machine Learning, Predictive Analytics, Neural Networks ... These are just some of the names used to describe some of the things that have been worked on since the 1950s. So why all this hype about AI now? What has changed?

AI has been around for the past 70 years, but the bulk of it has been secluded to the labs of research institutions. However, in the past decades, two things changed. We have experienced some impressive advancements in both the software and the hardware of AI systems.

Algorithms which have been around for 50 years and which had reached the limits of their capabilities were improved upon and a new breed of algorithms called Deep Learning emerged. These algorithms mimic more closely the internal workings of the brain. The improvement of these algorithms has been so profound that AI is managing to surpass human performance in most tests.

But these algorithms are extremely power-hungry and require huge processing power. During the same period when these algorithms were being developed, a new kind of processor was becoming popular. This processor is known as the Graphical Processing Unit (GPU) and (as its name suggests) it is designed to handle the graphics found in games. However, AI researchers realised that the GPU can be ideal for AI too and in fact, it is heavily being used for high-end processing. In some cases, the GPU can offer up to 100% speed-up over traditional processors. If that is not enough, we also have Cloud technology; which is essentially a farm of processors located somewhere remotely. The benefit of Cloud processing is that one can add or remove processors based on need. This combination of better algorithms and powerful hardware is currently driving this AI revolution which we are experiencing today.

With Malta’s limited R&D budget, what kind of AI is it possible to create in Malta?

Really and truly, we can create any kind of AI. Considering that most of the processing is performed on the Cloud, we are not limited by the lack of processing power, since we use remote machines. The advanced algorithms required for AI are all available and in fact, they are already taught at University. This means that we are already at a very advanced stage. We have projects in the manufacturing industry together with multinationals. Others are in health, together with different departments of Mater Dei. A lot of projects deal with the processing of big data and the prediction of future trends. We are constantly training AI experts in order to meet the demand which currently exists.

So we are positioned very well as a country. However, we can always do more. The R&D budget is very limited and needs to be boosted further. The government is aware that we can never compete with the big players like China which is investing around $100 billion on AI, but there are niche areas where our size plays in our favour. That is why the National AI Strategy focuses on the concept of the AI launchpad – because solutions are tested in a controlled sandbox, within the context of an entire country and then launched to the world.

Elon Musk, CEO of Tesla, said: “The pace of progress in artificial intelligence is incredibly fast ... You have no idea how fast it is growing; at a pace close to exponential. The risk of something seriously dangerous happening is in the five-year timeframe. 10 years at most.” ...Can artificial intelligence be dangerous?

AI is essentially a tool. To be precise, it is the most powerful tool ever invented by man. Like any other tool, it can be used to do good or to do really bad stuff. However, even though AI is progressing at a very fast pace, we are also aware of its limitations. In fact, the AI we have so far is normally referred to as Narrow AI. This means that the AI is exceptionally good at solving a particular problem but really bad at solving a generic one. Imagine, you’re playing chess against the AI. The chess program is probably at a level of a chess grandmaster or even better.

However, if you ask the chess program information about the weather, it will not give you an answer. The reason being that it is only an expert at playing chess and at nothing else. In fact, the holy grail of AI is what is known as Artificial General Intelligence (AGI) whereby the AI is capable of handling different situations in a similar way to humans. However, AGI is still very far away and it is very unlikely that we’ll see it in the near future.

Market leaders insist AI needs to become more broadly available in business. What are the benefits of introducing AI-powered solutions to a business today?

The benefits of AI to businesses is various. First of all, we have to remember that when we speak about AI, it is not just a program but rather a collection of different algorithms. AI is made up of different subfields too, each with its own suite of algorithms. They make the computers see, talk, learn and understand how the physical world works. Because of this, businesses can gain a lot of advantages from AI. It could be a simple matter of targeting specific customers when they publish their adverts, up to optimising hundreds of machines in order to improve their yield. AI will also bring forth better efficiency, quicker response to problems and make future predictions which would be otherwise impossible. AI will also provide new services to customers. Chatbots will be on standby, 24/7 ready to answer any query. They will also guide the user in buying online and shipping them almost immediately. For the first time, AI will assist businesses to provide a unique personalised experience centred around their clients.   

Is AI the playground of only large multi-nationals? Or can - and should - even small business look into how they can integrate AI in their service solutions?

No, AI is not just for the large multi-nationals. A lot of small businesses can benefit from AI and most of them are already doing it. Just think about Facebook or Google Ads. Business can set them up in a few seconds and let the AI decide when to show an advert. Today, many companies on Facebook use a chatbot. The chatbot is the AI that is handling customer queries when no one can answer such as when the shop is closed. A lot of small companies, such as barbershops, have an online system to schedule appointments, in some cases saving them 24 hour a month stuck to the phone and taking appointments. The AI is at the heart of such systems, improving the lives of people by automating simple repetitive time-consuming tasks.
Malta will definitely need to build AI skill-sets in the future workforce. But the potential for technological job displacement is also a major concern for policymakers. Will jobs in Malta be automated and will people be left behind?

The most critical resource, in any country, is human capital. A skilled workforce and a well-oiled educational system lie at the foundation of high-value-added services. With the proliferation of AI, we expect existent work practices to be disrupted, thus creating not only new opportunities but also new challenges.

We have seen a lot of predictions in the past year. Some experts foretell that around 40% of all jobs will disappear in the coming years. Others say that the AI revolution will create millions of new posts, some of which have not even been invented yet. There is nothing new in this. If we have a look at what happened when the banks introduced Automated Teller Machines (ATMs), we can see a similar pattern. With the deployment of ATMs, many predicted the end of bank tellers. However, statistics show otherwise. In 1985, there were 60,000 ATMs in the US and 485,000 bank tellers. In 2002, the number of ATMs rose to 352,000 while the number of bank tellers rose to 527,000. Our interpretation of this is that many people were finding it convenient to make use of the new machines; hence, the number of banking transactions began to soar.

On the other hand, banks started focusing on better customer service; thus adding more employees to their branches to handle more complex tasks. We can also see this pattern happening today with online banking. Even though some banks closed several branches and 90% of transactions are taking place online, the number of bank employees in the US is practically the same. It goes to prove that it is not just a matter of automation taking over human jobs.

The truth of what will happen is most probably somewhere in between. Let’s make it clear that in most cases, AI will not take over jobs, but it will automate specific tasks within that job. In essence, the job market will change as follows; Some jobs will become obsolete such as driving.

Self-driving cars will be able to transport people and goods around without the need of a human chauffeur. Some jobs will not be affected much by AI, such as nursing. In the caring profession, the human element is still critical and essential. Some new jobs will be created, such as Organ Creators, whose role will be to develop organs and body parts from organic material. Many existent jobs will change forever.

These range from low paid jobs such as store assistants up to high paid jobs such as lawyers. AI will augment their tasks in such a way as to make them safer, faster and more precise. To prepare for this revolution, we need to help the workers of today and those of tomorrow, adapt and acquire new skills.
Where do you see Malta in this AI world in 10 years’ time?

The coming decade will be a fascinating period for Malta since we will see the implementation of the new AI strategy. There’s a long way to go. One which will help us reduce our educational gaps, ensure that everyone achieves his full potential, support children at risk, re-skill our workforce and mould the professionals of the future all using AI.

It will not be easy, and the road ahead will take us towards exploring unchartered territory. What’s for sure is that we are in the right direction towards turning Malta into an AI launchpad. We are working not only to shape our island with the use of AI but I am confident that if we continue in this direction, we will also be an example of technological advancements for other countries in the coming years.

More in People