WE use artificial intelligence (AI) every day in our mobile devices and computers, and companies also use it to analyse data in their decision-making process. But the true potential of this technology has yet to be realised.
At least that's what Dame Wendy Hall, the regius professor of computer science and associate vice-president (International Engagement) at the University of Southampton Malaysia, believes.
"AI is everywhere, but it still has not become the norm today. The big tech companies use AI to analyse and monetise user data, but utilising AI for decision-making is still in its infancy in many sectors," she said.
AI has evolved over the last 30 to 40 years, from the early days of being an expert, knowledge-based and recommender system to the era of machine learning and big data that we are living in today.
AI machines are being trained to handle vast amounts of data, allowing their analyses to be incorporated into our daily lives in ways that we don't realise.
"A good example is facial recognition technology that has evolved by leaps and bounds since it was first developed. When I was a young lecturer, facial recognition in general was considered to be an unsolvable problem.
"Today, the power of AI allows machines to recognise faces, even when we have our face masks on. This becomes both exciting and dangerous as issues such as privacy violations are also on the rise," she said.
Today, most people will interact with some form of AI when using their smartphones, especially when they go online and access sites and apps, such as Google, Facebook, Instagram, Twitter and more.
However, despite its capabilities, AI is still far from perfect.
"Today, help centres are becoming increasingly automated. For example, we are seeing an increase in chatbots. Human operators have disappeared, with chatbots taking over the role of answering our questions and solving our problems. But we all know these chatbots are very limited in their responses, causing us to go round in circles and wait for hours to talk to an actual human being. They are not AI in the true sense. We still have a long way to go," explained Hall.
NEXT GENERATION OF AI
Hall, who has been in the technology industry since the 1980s and was part of the team who developed the Microcosm hypermedia system in the mid-1980s, which was a forerunner of the World Wide Web, said although AI is used widely, we have not achieved the level of Artificial General Intelligence (AGI).
"Machines are currently trained in a way that is limited to a specific sector or skill. For example, an AI that plays chess can beat a grandmaster, but it will not be able to do anything else or transfer those skills to other learning systems.
"On the contrary, we humans learn lots of things and we can link our knowledge across various skills. If we learn about one thing, our brain enables us to apply that to other things," she explained.
"If AGI is achieved, intelligent machines could free humans from monotonous work, allowing us to focus on higher value-adding work," she said.
But as humans, we do not like to be controlled or taken over by machines. We've seen in movies like The Terminator what could happen if machines take over. Even many philosophers argue that there is a risk of intelligent machines developing faster than humans, potentially resulting in the end of the human race.
Reacting to this, Hall said: "We must ensure that governance issues are thought through so we can use AI to our benefit instead of letting them 'take over the world'".
As for what the future holds, we need to ask — How do we still keep human beings in the loop?
"Soon, we need to think about how we can manage the development of AI systems to ensure that they do not take over the human race, but instead, complement our lives for the better," she added.
RESPONSIBLE AI
Hall is currently making headway in the development of responsible AI.
"It's about how we create human-centred AI systems. In other words, how do we design AI systems that ensure humans have been part of the entire process from design to testing, evaluation and operation going forward," she explained about her new project.
"If a company is using AI, it needs to have a process in place to make sure that the AI algorithms and data they are trained on are not biassed and are used ethically.
"It can be difficult to determine what is ethical and what isn't. We could have very high-level ethical principles claiming that AI will do no harm, but it is a far-fetched thing to achieve. It is difficult to define 'ethics' in this context, hence, I prefer explaining it from the perspective of 'responsible AI'," she said.
AI FOR MALAYSIA
Hall said the advent of AI does not mean that the future workforce will consist of only computer science graduates. Rather, AI skills in their broadest sense should be learnt by students in all disciplines to complement their work.
For instance, a student can major in business studies while taking an AI module as an elective subject.
"Malaysians who are already in the workforce can also leverage AI specialist training courses to further understand and embed AI technology into their respective industries.
"Plans are already in motion as the Malaysian government has spearheaded the development of the AI Roadmap (AIRmap) to establish AI governance and foster AI talent in the nation," she said.
AIRmap's creation was led by experts from Universiti Teknologi Malaysia and supported by leading industry consultants from Pikom, as well as the Science, Technology, and Innovation Ministry's National Science and Research Council.
"The bottom line is we need people to understand what AI is and how it works so that it can be applied in various industrial sectors. To achieve this, we need to get colleges and universities to start training and investing in the next generation now," she explained.
Hall, however, said Malaysia has a long way to go in AI development.
In the workplace, for example, McKinsey revealed that 50 per cent of work time in Malaysia is spent on repetitive activities that are highly automatable.
"AI can automate routine tasks, augment employees' capabilities, and allow time for us to focus on stimulating and higher value-added work. There is a lack of AI talent in the nation as it is difficult to find people trained in AI or having the skills to apply AI in various industries. Hence, we see the need for Malaysia to invest in AI skills training," said Hall.
"With UK universities having established AI strengths, there is the potential for UK and Malaysian universities to work hand-in-hand to build up Malaysian talent and introduce the local industries to the potential of using AI," she added.
A good sign is that the government has been very supportive of the endeavour. Under the 4IR and MyDigital blueprint, the government aims to boost the country's productivity by 30 per cent across all sectors by 2030, compared with 2020, with AI as a critical enabler.
WILL MACHINES REPLACE PEOPLE?
Hall says the situation in Malaysia may be different, though she imagines it is the same as everywhere else in the world. Post-Covid, there are shortages of staff everywhere, particularly in sectors that cannot deploy AI readily, such as restaurants and the services industry.
"While AI is able to take over mundane production or administrative tasks from humans, there are many jobs and necessary tasks for which robots and AI cannot replace human beings. We should be looking to use AI where it can make systems more efficient and productive, and free us up to undertake more creative roles and do the things only human beings can do," said Hall.
"It's about striking a balance between technology and society. I don't buy into the argument that says AI is going to take over all our jobs. Instead, technology will pave the way for the creation of more meaningful jobs in the future," she said.