6 minute read

ARTIFICIAL INTELLEGENCE

Next Article
AMADOU GALLO FALL

AMADOU GALLO FALL

ARTIFICIAL INTELLIGENCE

Ben West

The first time I was processed by a machine at passport control rather than by some jaded bloke looking me up and down suspiciously, it seemed a marvel. A machine that could assess whether my face matched the photo in my passport seemed terribly sophisticated.

However, I’ve just walked through passport control at Dubai Airport, and the machine didn’t even look at my passport. After five seconds my name came up on the screen, and I was let through.

This is just one of thousands of examples of artificial intelligence, or AI, becoming more and more complex. The pace of advancement is so fast now, we hardly bat an eyelid when we encounter a new incredible technological breakthrough. AI will certainly continue to change our lives profoundly in the next few decades.

So, what exactly is AI? Science fiction may paint it as robots with human-like characteristics, but in the real world AI can encompass anything from Google’s search algorithms to organising our personal and business calendars.

Apple’s virtual assistant SIRI and Amazon’s Alexa are AI, as are delivery robots, voice recognition and recommendation engines that suggest products you might like based on what you have bought in the past.

Further examples of AI applications include responding to simple customerservice queries, generating 3D models of the world from satellite imagery, or coordinating with other intelligent systems to carry out tasks like booking a hotel at a suitable time and location. The impressive tasks carried out by this technology multiplies each day.

AI systems demonstrate behaviours associated with human intelligence, which could include learning, reasoning, planning, problem-solving, perception, manipulation, knowledge representation, motion, social intelligence and creativity.

To learn, the AI system is fed large amounts of data, which it then uses to learn how to carry out a specific task, such as translating text or recognise a face.

As our understanding of AI advances, it is gradually moving from ‘narrow’ or ‘weak’ AI, where it is designed to perform a narrow task, such as internet searches or driving a car, to ‘strong’ or ‘general’ AI, where an intelligent agent is able to understand or learn any intellectual task that a human can, the eventual aim being to create intelligent machines that are indistinguishable from the human mind.

Narrow AI systems can only learn or be taught how to do defined tasks, while general AI has the type of adaptable intellect found in humans, a flexible form of intelligence that is capable of learning how to carry out very different tasks. These could be anything from creating spreadsheets to flower arranging or cutting hair.

While narrow AI may outperform humans at whatever the specific task is, such as solving equations or manufacturing a component, general AI would outperform humans at nearly every cognitive task. This kind of AI is typically seen in films and does not exist as yet. Estimates on how long before we get to that level of sophistication vary from a few years to centuries.

When modern AI research began in the mid-1950s, it was thought that we’d master general AI quite quickly. Indeed, AI pioneer Herbert A. Simon wrote in 1965: “machines will be capable, within twenty years, of doing any work a man can do.” Such conjecture triggered a spate of wild sci-fi novel and film creations from the likes of Arthur C. Clarke and Stanley Kubrick. However, by the 1990s it was becoming apparent that the AI journey was going to take much longer than first thought.

However, the field is definitely advancing. For example, in 2017, one study that conducted intelligence tests on publicly available narrow AI such as SIRI and Google AI found that they reached an IQ value of around 47, which corresponds approximately to a six-year-old child. An average adult has an IQ of about 100. Tests carried out just three years before found that programmes like SIRI reached a maximum IQ value of 27.

Breakthroughs have been mounting up relentlessly, whether they’re computer systems beating chess grandmasters, or the best players on US quiz show Jeopardy!.

Roll on to 2020 and AI research laboratory OpenAI developed Generative Pre-trained Transformer 3 (GPT-3), which uses deep learning (ie machine learning) to produce human-like text. Although it is generally not considered to be an example of general AI, also it is too advanced to be seen simply as a narrow AI system.

The quality of the text it generates can be so high that it can be difficult to determine whether or not it was written by a human. In one regard, this is very encouraging, but in another, it is a cause for concern, as it has the capability of spreading misinformation, creating fraudulent academic essays, phishing, abusing governmental and legal processes and other dangerous behaviours.

However, further testing found that many sentences generated weren’t of a particularly high standard at all, that there’s still a long way to go for a machine to compete fully with a human.

More generally, AI opens up potential for numerous threats. For example, the technology is close to being able to replicate a person’s voice perfectly, and create photo-realistic images. It may cause profound disruption, as the inevitable conclusion is that we will no longer be able to trust video or audio footage as genuine. Facial recognition may increasingly be used for surveillance by some territories to the point that privacy is all but completely lost.

There are fears that AI will take away many jobs in the workforce. The more routine and repetitive, the more at risk a job of being swallowed up by AI. But with AI abilities improving so much, before long all jobs could be threatened, whether lorry drivers, journalists, lawyers or surgeons.

There is disagreement over how much of a threat AI is to human existence. SpaceX and Tesla CEO Elon Musk believes that AI is a “fundamental risk to the existence of human civilisation,” while physicist Stephen Hawking warned that if AI got too advanced, it would quickly greatly outstrip human capabilities and could pose an existential threat to the human race. However, a number of prominent people working in the field of AI disagree that there’s such a threat.

AI certainly does open up many questions for the future. Will our employment vanish, or enough new work be created? Would it be advantageous for there to be no jobs and where instead everyone enjoys a life of leisure and machine-produced wealth? If we create super-intelligent machines, will we control them or will they control us? What will it mean to be human in the age of sophisticated artificial intelligence? Whatever the answers, we’re certainly entering interesting times!

THE METAVERSE

One huge buzzword of late is the Metaverse, an online virtual world that incorporates 3D holographic avatars, video, augmented reality, virtual reality and other forms of communication. Ultimately it will offer a hyper-real alternative world to coexist in, with participants playing, working and staying connected with friends through such things as conferences, live events and virtual travel trips.

The metaverse is enabled, populated by and supported by AI. It was first coined in 1992 in Neal Stephenson’s sci-fi novel Snow Crash and companies such as Microsoft and Meta (formerly Facebook), Minecraft and Fortnite have wholeheartedly embraced it. As well as creating an augmented reality for playing online games, with users wearing virtual reality headsets, it can add another dimension to such activities as socialising and business meetings. However, it has been criticised for such things as creating an arena for encouraging grooming and harassment, especially as it is not yet suitably regulated.

As it increasingly takes hold, society is less and less considering the virtual world to be fake, but instead is a new state of human existence.

This article is from: