In 1965, philosophy professor Hubert Dreyfus criticised AI for its limited uses, pointing out that computers couldn’t even play a decent game of chess. Two years later, he was checkmated by a computer. Read on to find out what artificial intelligence is, what it can do (aside from play chess), and whether it’s likely to help or harm society.
What it is
Artificial intelligence (AI) is technology that enables a computer to do things we usually associate with human intelligence. This includes:
- learning from past experience
- adapting to new circumstances
- making generalisations
- solving problems
- making decisions
A short history
In the 1940s, a young mathematician called Alan Turing wanted to find out if it was possible for machines to think. He reasoned that humans can make decisions based on information and logic, so why couldn’t machines? He published a groundbreaking paper called ‘Computing Machinery and Intelligence’, where he set out his ideas on how to create intelligent machines and test their intelligence: if a machine can have a conversation with a human without being detected as a machine, it has shown human intelligence.
So far, nobody’s made a machine that can pass the so-called ‘Turing test’, but we have made some big advances:
- In 1966, Shakey became the first robot to perform multi-step actions based on reason. If it was told to ‘push the block off the platform’, for example, it could look around to see which platform had a block on it, find the ramp needed to reach the platform, move it to the platform, roll up the ramp, and push the block off.
- In 1977, Deep Blue defeated the world chess champion, Garry Kasparov, by quickly calculating every possible move and the outcome of each one.
- In 2015, computers outperformed humans for the first time in the annual ImageNet challenge, where they correctly identified more than 97% of the images displayed.
Today, AI is everywhere: it’s used to diagnose medical problems, recommend places to eat, and even predict natural disasters. And as we get better at collecting information and technology becomes more advanced, it’s likely that AI will play an even larger role in our lives.
Why more AI could be a good thing
Less human error
We often use facts and statistics (called ‘data’) to make important decisions - which route a plane should follow, for example, or how to prevent a disease from spreading. But when humans collect lots of data, we usually make mistakes. These mistakes, often caused by things like tiredness or lack of attention, are called ‘human error’, and they can have disastrous consequences.
AI can cut back on human error by breaking down huge amounts of data for us. And because AI doesn’t make mistakes if it’s programmed correctly, it can be trusted with some pretty big decisions - the US Navy, for example, uses AI to predict part failures within its ships and plan preventative maintenance.
We can use AI machines in very dangerous situations - exploring the deep ocean, for instance, or defusing a bomb. That’s because purpose-built machines have a much better chance of surviving hostile atmospheres than humans, and even if they don’t survive, it’s better than risking human life.
As an example, researchers are trying to develop a robot that can put out fires by making autonomous decisions - things like where the fire is coming from, or how to position its water hose most effectively. With climate change contributing to 10,000 daily wildfires worldwide, this kind of robot could help to save many human lives in the future.
Most people don’t work well all day long - research has shown that 60-80% of us work best in the morning, and long working hours have been linked to health problems like cardiovascular disorder and headaches.
In contrast, AI machines can work without breaks and keep a high level of productivity. This could be very useful in roles that are needed 24/7, such as online customer service for an international bank, or healthcare for the elderly.
Why more AI could be a bad thing
Fewer employment opportunities
A common worry about AI is the impact it could have on employment. We seem to have adjusted to machines replacing people for fairly simple tasks like scanning items at a till, but thanks to AI (and progress in other areas like robotics), machines can now do some of the tasks of doctors, architects, and even artists. This could make it harder for people to find work in these areas.
A study carried out by the McKinsey Global Institute estimated that up to 30% of the hours worked globally could be automated by 2030. The same study pointed out, however, that this might not cause a loss of jobs, but rather a demand for different skills. The study suggests that less ‘predictable’ jobs like gardening and plumbing will be less affected, and the demand for other jobs, such as managers and creative professionals, might even increase.
Not enough compassion
The late Joseph Weizenbaum, a computer scientist and professor at MIT, argued that AI technology shouldn’t be used to replace people in jobs that require compassion - jobs like therapist, judge, or nurse. This is because AI is incapable of having genuine empathy for others and would struggle with, for example, helping a patient to cope with some bad news about their health.
When someone does something that harms another person, we can use the law to figure out who was responsible and what their punishment should be. With AI, things are a little more complicated.
Let’s take self-driving cars, for example. If a self-driving car powered by AI crashes into another car and hurts the person inside it, who would be to blame? The person who bought the car, the car manufacturer, or the people who programmed the AI software? Without an obvious individual to blame, we’ll need clear laws that can tell us who is legally responsible.
So, what do you think about artificial intelligence? Do you see it as a valuable tool to make our lives easier, or do you think compassion (and legal responsibility) is essential for making big decisions? If you’re keen to learn more, take a look at the links below.
Good stuff from elsewhere
The Turing Test - Britannica
Find out more about Alan Turing and the beginning of AI.
AI used to complete a Rembrandt painting - ARTNews
An article about how a famous damaged painting - The Night Watch, by Rembrandt - was successfully restored using artificial intelligence.
McKinsey report on the future of automation
A report that looks at the possible loss of jobs in the future thanks to automation, and the demand for new jobs.