About AI technology

Your guide to everything AI

Artificial Intelligence – or AI – is the next frontier in technologyIn its most basic terms, AI is the simulation of human intelligence processes by machines. It has the potential to transform our society for the better with countless applications.

However, there are serious concerns about what could happen, if left unchecked, about the untold damage AI could cause. 

In this article we’ll define what we think AI is, discuss some of its applications as well as present arguments from both sides of the debate about its advantages and disadvantages and discuss its future. 

A brief overview of Artificial Intelligence 

For clarity, we define Artificial Intelligence as: 

The simulation of human intelligence in machines or computers. This includes learning, reasoning and self-correction.  

AI is an industry term that’s started off as a science fiction fantasy, but in recent years has become an integral part of day-to-day life. The field was initially founded in the 1950s by scientists John McCarthy, Marvin Minksy and others.

Unfortunately, AI didn’t take off immediately, primarily because of the lack of computational power and memory available to scientists in the middle of the 20th century.  

It wasn’t until later in the 20th century, a renewed focus on AI brought some of the innovations we enjoy today. In 1997, Deep Blue, an artificial intelligence created by IBM, beat chess grandmaster Garry Kasparov in a six-game match 3 ½ – 2 ½.

Deep Blue became the first computer to defeat a reigning chess world champion. 

There have been arguments made by commentators that Deep Blue passed the Turing Test, a methodology developed by Alan Turing, to identify whether an AI is a true artificial intelligence.  

In very simplistic termsa pass for the Turing Test is centred around whether the AI in question can convince a human that it is a human itself. And that’s what some say Deep Blue managed to achieve when Kasparov accused IBM of cheating after he had been defeated. 

He demanded a rematch because he thought IBM was using a human opponent – thus Deep Blue passing the Turing Test.   

Deep Blue marked a significant turning point in public perception of AI and its applications. Popular media has overwhelmingly portrayed AI as an antagonistic entity, hell-bent on global domination and destruction. .

Films such as The Matrix and Terminator 2: Judgment Day certainly didn’t help this perception, but indeed, it could be argued that Deep Blue marked the start of public awareness of the positive things AI could do for humankind. 

As the world continues to hurtle through the 21st century, AI innovations are happening at a breakneck pace. Since 2000, the world has seen: 

  • Smart toys; self-enclosed novelty items that can leverage cloud-computing to respond to perform rudimentary tasks 
  • Mars ROVERS; autonomous, AI-driven machines exploring the surface of Mars and reporting information back to NASA 
  • Recommendation technology; algorithms that learns and understands your habits, interests and desires and makes recommendations based on them. This technology underpins social media, internet searching and thousands of other applications 
  • Autonomous cars; driverless cars that operate with an AI at the helm instead of a human being 
  • A series of game-driven AI’s  specific AI’s have been developed to challenge and beat the best players of games such as Jeopardy, League of Legends, Poker and Go; one of the most complex games ever invented by humans. 

Types of Artificial Intelligence Collection of robots

There are four types of AI according to Arend Hintzea researcher at Michigan State UniversityThese are: 

Type one – reactive machines  reactive machines can only react to what is front of them, it has no memory and can’t use past experiences to advise future ones. Deep Blue is an example of a reactive machine; it recognised chess pieces and that it was indeed playing chess but used its computational grunt to make the most strategic decisions based on state of the board as it made its move. Reactive machines are designed for extremely narrow purposes and can’t be implemented easily elsewhere. 

Note: This is why other commentators argue that Deep Blue didn’t pass the Turing Test, because it isn’t a true AI, with the actual ability to deceive Kasparov – it’s all a matter of interpretation.  

Type two – limited memory –this AI uses previous experiences to inform future decisions. For instance, driverless cars are designed to observe other cars and how they move around them to make decisions.

Coupled with programming like reactive machines, whereby cars will always brake if a person runs out in front of the car, driverless vehicles learn as they experience more, but without permanently remembering learned experiences. 

Type three  theory of mind – this system understands different emotions and will be able to differentiate between feelings and respond accordingly. This type of AI is still under development as we haven’t quite perfected the technology yet. 

Type four  self-awareness – AI’s that demonstrate self-awareness, thought and have their own consciousness are the pinnacle of artificial intelligence. They understand their own current state and use the information to infer how other people or machines are feeling.

This type of AI is the most likely of all to pass the Turing Test, as when the technology is invented, speaking to this AI will be indistinguishable from a human being. 

How is AI used right now? 

Right now, AI has a huge range of applications in your day-to-day. We’ve prepared this SlideShare with six of the most important, which you can see below:  

Arguments in support of AI 

Artificial Intelligence is a hotly contested subject, with the world’s leading minds offering conflicting views on the technology.  

Jack Ma, a Chinese entrepreneur who until recently, headed up Alibaba Group, thinks that AI will be helpful to humanity. Ma also thinks that machines will never replace humans because they’ll never be as good as humans at certain tasks.

Recently, Ma took to the stage in China and debated fellow entrepreneur Elon Musk about the merits of AI. The highlights are enclosed below:  

Jack Ma’s sentiments are part of wider-narrative that AI is set to liberate humanity from tedious tasks and allow our society to fundamentally reconfigure. 

In essence, AI’s supporters say that machines will help us work less and focus on more creative tasks which humans are more suited to. 

Broadly speaking, arguments in favour of AI can be broken down into these four categories: 

Eliminates mundane tasks – data entry, data analysis or other boring tasks can be automated by AI technology.

This frees up precious time and resources for human beings to focus on creative pursuits, something that AI can’t do right now. Examples of tasks that AI can automate include 

Accelerated decision making  AI is capable of analysing reams of data from disparate sources to make decisions quicker.

This can help humans make important decisions without having to spend brainpower analysing data required to make a choice.  

Eliminates errors – an AI is never tired, distracted and is consistent in its approach to everything it ever does.

It doesn’t make mistakes, because mistakes aren’t part of its programming, assuming of course it is programmed properly! In other words, vast datasets can be processed quickly, efficiently and in a fraction of the time without error. 

Reduces human risks  one of the most important facets of AI is that it empowers machinery to take risks on behalf of humans. For instance, accidents can be eliminated from productions lines, as machines carry out complex tasks usually assigned to humans.

In addition, extra-terrestrial exploration such as activities on Mars, are all carried out by AI-driven walkers, eliminating the need to send human beings on hazardous expeditions. 

Arguments against AI 

While AI does offer compelling opportunities to enhance our society, there are downsides which are hotly debated by the scientific and business community. Professor Stephen Hawking, one of the world’s greatest scientists, issued a stark warning about the dangers of AI in 2014.

The professor said, ‘AI could spell the end of the human race’. 

His anxieties centred around what would happen when humanity developed a true, sentient AI and the exponential rate it could learn, develop and change itself. Hawking’s issues are detailed in this lengthy blog post by Tim Urban, who discusses AI and the concept of recursive self-improvement. 

Recursive self-improvement is a mechanism used by AI researchers to incrementally improve the performance of an AI for a specific scenario. A high-profile example of this is the Dota 2 experiment run by Elon Musk’s Open AI.

It beat a top Dota 2 player, by playing thousands of games against itself in six months to get to the same level as its human counterpart. 

The losing player remarked: 

“It had learned in six months, what I had learned in eight years.” 

This anxiety underpins much of the concern the scientific and wider community has about AI. If an AI had sentience, how quickly would it overtake humans as the smartest entity in known existence?  

And what would its intentions be? 

Other concerns about AI include: 

Very high cost  according to the Wall Street Journal global AI spending is set to hit $35.8 billion in 2019.

This vast sum is largely down to costs for maintaining, altering and improving AI incrementally. The technology is still under development and as such, research needs to be carried out to steadily improve it. 

You can’t replicate human decision-making; for now  machines don’t have emotions or any sort of moral values. They execute instructions based on their programming, with no perception of what is right or what is wrong.

On a small-scale, this might involve rejecting a credit application based on criteria not being met. AI doesn’t consider extenuating circumstances and sees the world in black and white.  

Machines lack creativity  while AI can help you design and create things, as it stands, they can’t match the creativity or power of human thinking. 

With previously cited examples such as Deep Blue and Open AI, both these AI’s can become specialists at what they do through recursive self-improvement but lack the ability to apply themselves to other tasks. They simply cannot consider variables outside the parameters of their programming. 

Mass unemployment – almost half of the world’s jobs are vulnerable to automation according to a landmark study issued by OCED. 

The most susceptible jobs include those in warehouses and manufacturing, data entry and analyst roles. Automation will cause a significant period of social upheaval where the skills that workers have for jobs that have been automated will become redundant.

Some analysts predict mass unemployment and social unrest. 

IA could be the future of AI 

Recursive self-improvement is one key advantage AI has over human beings. It allows an AI to rapidly learn from its own errors, at a rate that would take humans hundreds of thousands of years. 

But what if there was a way to level the playing field? What if there was a way to effectively, hack evolution and progress at an exponential rate like AI? The answer could be IA. 

Intelligence Augmentation (IA) is technology designed to complement human intellect.

Effectively, it involves using technology to augment or enhance human intelligence. While it might sound like science fiction, it is science fact, as Elon Musk announced he’s initiated a project to develop something called a neural lace; an interface that connects data directly to the human brain. 

You see him talk about it with popular podcaster Joe Rogan. 

Musk is effectively proposing technology that allows the wielder to work with AI and process information quicker. Technology has facilitated the creation of more data in the last two years than the entire history of humanity. 

But this means humans are bombarded with more information than ever, with only their fingertips to process it. Neuralink aims to speed that process up by using AI to help humans think quicker. 

In a presentation this year, Musk and his team at Neuralink revealed that preliminary trials and animal testing have yielded some success. They successfully implanted a Neuralink in a chimpanzee that allowed it to control a computer with its mind. 

In 2020, Neuralink hopes to have their first human trial; subject to approval from medical authorities. 

The potential applications for Neuralink are staggering. While the research is at a preliminary stage and hasn’t been tested yet on humans, it could fundamentally change our society and species.  

Recipients would benefit from the assistance of technology to help them process information quicker, without having to physically move their fingertips. It could also transform data processing, working and a whole host of other things we haven’t even thought about. 

However, Neuralink comes with grave ethical concerns. Would humanity cease to exist as it is? Would it give recipients an unfair advantage over non-recipients? 

What do we think about AI?

What do we think of AI?

We think AI is a valuable technology for our society. It has the potential to elevate humanity in ways we haven’t experienced since the Industrial Revolution or the Dot Com Boom.

It will help liberate us from some of the laborious and repetitive tasks that we clearly aren’t as good as machines at. 

However, we think there will be serious upheaval and fundamental change required to accommodate it.

And it will require the combined effort of politicians and businesses to appropriately prepare the world and its people for the changes that are inevitably coming. 

It’s imperative that we learn from the past to ensure that as we digitally transform, again, we ensure that nobody is left behind and is given the opportunity to put their talents to use. 

What do you think about AI?  Share your thoughts by visiting our  LinkedIn or Twitter pages and letting us know. 

Resources 

We consulted a range of resources to produce this content and we recommend you read through them if you’d like to find out more. 

Books 

  1. The Singularity is Near (the): When Humans Transcend Biology – By Ray Kurzweil
  2. Superintelligence, Paths Dangers, Strategies – By Nick Bostrom
  3. Rise of the Robots – by Martin Ford

Videos 

  1. Machine learning – living in an age of AI – https://www.youtube.com/watch?v=ZJixNvx9BAc&t=3s
  2. How AI can save our humanity – Kai-Fu Lee https://www.youtube.com/watch?v=ajGgd9Ld-Wc
  3. Google Duplex – https://www.youtube.com/watch?v=D5VN56jQMWM&t=130s
  4. Elon Musk’s last warning about AI – https://www.youtube.com/watch?v=B-Osn1gMNtw
  5. Multi-agent hide and seek – https://www.youtube.com/watch?v=kopoLzvh5jY

Webpages 

  1. The AI Revolution: The Road to SuperIntelligence – https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html
  2. The AI Revolution: Our Immortality or Extinction – https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html
  3. What is AI?: https://www.techopedia.com/definition/190/artificial-intelligence-ai