Artificial Intelligence is one of the ‘hot topics’ in science; recently, Tesla’s Elon Musk announced he was beginning a new venture, Neuralink, to “merge the human brain with AI”. But apart from visions of cyborgs dancing the heads of science fiction writers, what are the implications of Artificial Intelligence? For the general public? For researchers? And for the future of employment? To get greater insight on this subject, we spoke to Professor Paul Chung of Loughborough University. Paul has been working in the field of Artificial Intelligence since the 1980’s. We began by asking him:
What inspired you to study Artificial Intelligence?
I did my first degree at Imperial College in computer science; in my final year, I took a module in Logic Programming which was taught by Professor Robert Kowalski. It was fascinating because it was so different to other things I’d learned; it featured a declarative form of programming that allowed the system to “reason” and provide its own conclusions. Indeed, this was so fascinating for me that I wrote a connection graph theorem prover in Prolog for my final year project ; I then went on to study Artificial Intelligence at the University of Edinburgh.
I remain fascinated by systems that permit you to say things at a high level of abstraction and then allow systems to draw their own conclusions.
What were your initial main challenges when you began to study Artificial Intelligence?
In the 1980s, a lot of the challenges were around knowledge representation, that is, how to present knowledge in such a manner that the system could carry out inference with it. Another challenge was trying to figure out how people learn; this was a major influence on the research at the time.
Knowledge representation has been superseded by ontology, which states the properties of things and the relationships between the things. In essence, it is a kind of knowledge representation that supports different kinds of reasoning. Research has moved towards machine learning; whether machines learn exactly like a human has been deemed less relevant.
How has research progressed since you began exploring the subject?
There’s been a tremendous take up of AI in conventional computing; it’s not usually stated “I’m using AI”, nevertheless, it’s embedded in their systems. People almost take it for granted. That’s a common theme: once you’ve invented an AI application, it becomes so commonplace that people take it as given. Furthermore, it’s more pervasive: it’s in engineering, medicine and other applications.
"...it’s not usually stated “I’m using AI”, nevertheless, it’s embedded in their systems. People almost take it for granted."
Let me further describe this with an analogy: it’s like motor technology. When motors were first invented, people said how great they were; now motors are everywhere and part of so many applications that we rarely stop and think about it. AI is very similar, it’s now part of larger systems. They are progressively becoming more intelligent and usable.
What are the current trends in Artificial Intelligence research?
The aim of AI has not changed: it’s always been about creating intelligent systems. Two of the main trends are machine learning and autonomous systems. Researchers have recognised the limitation of knowledge based systems: they are very rigid and it’s difficult to maintain their knowledge base. If the machine is able to learn from the information it receives, it eases maintenance. There is an ongoing fascination with how machines can go from knowing nothing and learning vast amounts.
As for autonomous systems: people want machines to be able to do things like mow the lawn. To build an autonomous system which can successfully move in a less restricted environment, it needs to be able to respond to unforeseen circumstances. This is also true for autonomous vehicles. These create new challenges and new research problems.
"To build an autonomous system which can successfully move in a less restricted environment, it needs to be able to respond to unforeseen circumstances."
Artificial intelligence is bringing different researchers together; you can’t have just the “intelligence”, you need to bring a variety of systems together, such as sensor technology, into a single solution.
Why do you think these trends have taken hold?
Mainly, it’s curiosity. People are fascinated with trying to build systems that have more and more intelligent behaviour. Furthermore, we’re trying to expand their scope. The challenge for researchers to build systems that can respond to real life situations like humans. This is supremely difficult; but it’s the challenge associated with the task that makes people want to pursue it. Think about it, a simple instruction such as ‘Please make me a cup of tea.’ requires a robot to do a great deal of reasoning and knowing quite a bit about the task such as planning the path of moving from its current position to the kitchen, knowing where the ingredients such as tea bags and milk are stored, how to fill and boil the water, etc.
What do you consider to be the most exciting and promising areas for research in this field?
It’s very difficult to tell at this stage; it’s like asking, where would you put your money? I believe that the most promising applications have to be ones with wide appeal. Google is so successful because everyone uses it. They are making it ever cleverer: ranking is becoming ever more accurate. The car is another potential example: at the moment, people have a fear of self-driving cars, wondering if they’re reliable and safe. However, the manufacturers would like to remove this concern.
Big Data will continue to make use of AI technologies: companies have a lot of data, and they want to make good use of it.
I think AI in games will be very popular: with AI, you can make the game more interesting and responsive to human beings in the way they are playing.
I also think the application of AI in medicine and engineering is exciting and will bring benefits to humanity such as new drug discovery and assisted living technologies.
Will AI facilitate or hinder serendipity in research?
I think it helps; researchers can use AI to help them advance their own research even if they are not computer scientists. In medicine, researchers can use AI go through the latest papers, find trends, and create summaries. Researchers in other domains can also use AI to help them. Furthermore, in order to create intelligent systems you need to be able to bring different disciplines together; it can serve as a catalyst for different disciplines – e.g., doctors, engineers, computer scientists – to work together. That will advance research. It’s a great opportunity to create multidisciplinary research projects.
"...researchers can use AI to help them advance their own research even if they are not computer scientists."
What parts of artificial intelligence are of most interest to government? Industry?
Data mining is important to governments: they have a huge amount of data, to analyse it manually would be very time consuming. They can adapt these technologies to improve the efficiency of data analysis. Data security is also vital application; when looking for unusual patterns in data, AI can be helpful. The military is already using drones in hazardous and dangerous situations: sending an autonomous robot will save lives. Autonomous robots will also be useful in rescue situations: there are places that are hard to get to: one can send a robot in for a search mission before a rescue.
As for industry, business analytics will remain important. Firms are always trying to understand the market and customer behaviour, and to encourage loyalty. Also: AI will help us understand trends. This in turn will enable companies to identify customer needs, create new products, and discover other useful areas that can be exploited. Intelligent manufacturing and diagnostic systems will also be very useful for industry. Another area in which industry can utilise AI more is safety: specifically, answer the question, how we can improve safety of operations? In hazardous process plants and so on, how can we improve their safety in the design stage as well as the operations stage?
How do you think artificial intelligence will affect the future of employment?
Some of today’s jobs are already being replaced by typical information technology. Again, it’s difficult to answer this question: let’s imagine if everything can be done by machines, or a lot of things can be done by machines: this leaves us with the question, what will the people do? What happens in a world in which even a machine can provide us entertainment content as well as do manual labour? What would the humans do? In my opinion, it comes down to services which require caring: people need human interaction, and humans need to take responsibility for decisions. The difficulty is who will get the jobs? And what happens to those whose jobs have been replaced? I can see a future in which big companies with all these machines will have more and more power. How will finance be distributed so that people are not impoverished rather than all the profits going to the very big companies that produce these machines? This remains a big question that has yet to be answered.
How do you think artificial intelligence will affect the future of research?
Researchers will have more tools at hand to do their research. Researchers are already benefiting from using search engines to find information. They’ll have a lot more advanced technology in their hands to help them do their research; data mining will help them find interesting patterns; various AI techniques can help them solve optimisation problems.
What should researchers and non-researchers do to prepare for these changes?
Change is inevitable; we saw this with word processors replacing typewriters. People use the ATM instead of going into the bank. As I said before; AI will be like the motor, it will just be part of everything we see. It won’t land suddenly; it will be pervasive. It is already in our phone; by and large it will help us with our tasks. It already helps us in doing our daily jobs.
The problem is when it becomes so pervasive that an autonomous system replaces a lot of the functions of a human being. Let’s again take the example of the car: we already have AI in it, to help us navigate, etc. Bit by bit, AI has arrived there, it even helps me parallel park; the only thing it doesn’t do is drive. We accept this is normal. For the moment, however, we retain ultimate control of the car, the technology is merely there to help us. But will it become so pervasive that it will be forced upon us?
Any additional words for our readers about the future of Artificial Intelligence?
If the aim is merely to make machines smarter and smaller, then there is no cause to worry about Artificial Intelligence. It’s just making our lives more convenient. If we’re creating machines to replace human beings, if we’re building systems that make more and more decisions for us, then I’ll be worried. As this suggests, the key question is: can the machine override our judgements? I’ll be worried if that ever becomes so: the machine may decide it doesn’t need our opinion any longer.
Thank you for your time.
How will Artificial Intelligence impact the general public and the researcher?
See our related jobs
|We found all these Artificial Intelligence Jobs||See every Intelligent Systems Job||Have a look at all our Machine Learning Jobs|