Is it AI or AGI? Nikolas Kairinos explains

103 0

Recent advances in the world of AI may be impressive, but just wait until you see what artificial general intelligence (or AGI) can do, writes Nikolas Kairinos, CEO and founder of AI think tank Fountech.ai


The proliferation of artificial intelligence (AI) has prompted as much excitement as it has concern over the past few years. On the one hand, many suggest the technology could solve society’s most pressing problems, reduce the amount of labour-intensive work that humans do, and generally make our lives a whole lot simpler. Others suggest the contrary: that it could lead to the demise of humanity.

As we make advancements in the field of AI, the subcategories of the technology will also grow, in the same way that the creation of moving images led to film, television and digital video. One variety that is of burgeoning interest to researchers and observers is known as artificial general intelligence, or AGI.

AGI and its impact

AGI differs from AI in that it refers to a machine or computer that has the capability of understanding the world as a human would. It doesn’t yet exist, of course, but the idea of it has seeped into our cultural consciousness via popular media.

Films like The Terminator and The Matrix showcase man-made robots with human-level understanding – but the real-world version of AGI will probably be much more quotidian. However, making any kind of solid predictions is a fool’s errand; although the field is evolving quickly, we are still far away from achieving something close to AGI.

For now, the real-world applications of regular AI are already being seen. Social media provides an excellent example; here, deep learning algorithms expose users to content that prompts the optimal emotive response, helping to create dependency and increasing ‘stickiness’.

While the impact that AGI will have is less clear, there is certainly huge potential in this field. Indeed, the possibility of a programme being able to process information as quickly as a human, all the while possessing the ingenuity of a person, would undoubtedly improve our decision-making abilities on both an individual and societal level.

Advertisement
Advertisement

In turn, this would free up time for people to divert to other activities. While some may associate ‘free time’ with unemployment, I consider it to mean greater time for an individual to pursue more value-adding and fulfilling activities.

In saying that, such a scenario would mean that AI would need to merge with human intelligence.

This would be functionally difficult, as it would require a programme to be augmented by millions of AI experts, to ensure human qualities are ‘coded’ into the AI. In order to achieve the best outcomes, machines will need to have the ability to present different options for humans to choose from.

With time, the AI will learn how to make similar decisions in a way that mimics human emotion and critical thinking.

The development of AGI presents risks, too. It would increase the ability of governments to conduct surveillance on their populations, as machines would be able to process a far larger amount of information — in the same way that Tesla cars are able to digest, in (almost) real time, masses of data about their surroundings and make movement decisions based on them.

On the flipside of the argument, Chris Bishop, laboratory director at Microsoft Research Cambridge, claims that these kinds of predictions are “utter nonsense”. At best, he argues, “such discussions are decades away.”

Dystopian concerns aside, then, AGI could help humanity combat hugely complex problems such as climate change. With a cohort of reasoning and high-functioning machines, our ability to analyse and deduce solutions will be massively increased.

How close is AGI today?

There is a growing body of evidence that we are already getting closer to achieving something akin to AGI.

Google Deepmind’s AlphaGo, for example, successfully beat champion Lee Sedol at board game Go. To put this in perspective, Sedol was once described as the “Roger Federer of Go”.

What was so interesting about this victory was that a computer is unable to calculate all permeations of Go – there are simply too many. So AlphaGo had to use reason and something bordering on intuition in order to win. This signified a major advancement on the road to genuine general AI.

The development of AGI will require greater creative thinking. We will need to see experts looking beyond reprogramming AI and instead considering new and novel ways in which AI technology can be applied

However, whilst Google Deepmind serves as an excellent example of just how far we have come in the development of AI, the technology remains just that: AI.

In other words, it still works by performing a limited series of functions, and lacks the key human qualities that would be required for it to be awarded the title of AGI.

Indeed, I fear that we are approaching an AI ‘winter’. While we cannot deny impressive breakthroughs like the development of deep learning – the science that underpins many advanced applications of AI toolsets today – the amount of research being dedicated to novel applications of AI has plateaued.

Programmes like AlphaGo might lead people to believe that AI can be ‘creative’, as might the ability of AI to create ‘deepfakes’, but these accomplishments do not yet meet human levels of creativity. True creativity cannot come from a system that solely takes input, performs analytical functions and presents the output.

The development of AGI will require greater creative thinking. We will need to see experts looking beyond reprogramming AI and considering new and novel ways it might be applied.

That said, this does not mean we have nothing to be excited about in the field of AI. There is little doubt that, as we progress toward achieving AGI, there will be major changes to how human society will be constructed. I, for one, look forward to seeing what the next big development will be.


Get in touch:AGI Nikolas Kairinos

Nikolas Kairinos is the chief executive officer and founder of Fountech.ai, an umbrella company to three specialist firms: Fountech.Solutions, Fountech.Ventures and Fountech.Science.

Fountech.ai is driving innovation in the AI sector, helping consumers, businesses and governments understand how this technology is making the world a better place.

Reach him here: @NickKairinos or @fountechAI

 


Leave a comment