They’re technologies which are already changing our everyday lives and will continue to transform the world as we know it. But the two terms Artificial Intelligence (AI) and machine learning are often seen as interchangeable when actually they mean two different things – which may not be that important to the general public but is when you’re trying to create a new service or tech product. So, here’s a quick reminder.
AI is the intelligence that is developed for machines, as opposed to the natural intelligence humans have. As the technology revolution has gained momentum creating artificial intelligence has become a huge, specialised branch of computer science.
It’s a broad term. Oxford Reference defines it as “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.” AI requires huge amounts of information in order to make sense of the world. To be described as such a system needs to be able to reason, problem-solve, understand and plan ahead. Depending on the tasks it’s been built for, it may also need to be able to see, hear and understand speech.
All of us who live amongst modern technology come into contact with AI in everyday life. Decades ago, we may have imagined a future of humanoid robots powered by AI – butlers who wait on us, metal hand and foot. But though they do of course exist, they haven’t gone mainstream. We haven’t yet achieved the multi-tasking companion-advisor androids of science fiction. Instead what we usually encounter is faceless but extremely useful AI built into the services we use every day, which react intelligently and problem-solve to assist us in reaching our goals.
Many forms of AI use machine learning: a much narrower subset of the field. It’s defined in the Merriam-Webster dictionary as ‘the process by which a computer is able to improve its own performance…by continuously incorporating new data into an existing statistical model”. Machine learning was founded as a field of study in the 1950s when scientists began talking about building an ‘electronic brain’, but it didn’t gain much traction until the technology advances of the late 1990s. Today, though it’s not a prerequisite of AI, machine learning is key to making it better, smarter and more sophisticated, in the same way that the ability to learn from our successes and failures is fundamental to our own, human intelligence.
Take the example of a chatbot. A basic chatbot is a form of AI that answers queries by looking for keywords and patterns against a database of knowledge and responding based on certain rules. Many won’t have been created with the capabilities for machine learning. But a chatbot or other form of AI using machine learning will be able to ‘learn’ from all the thousands or millions of interactions it has with humans, automatically getting better and better at things like understanding human language and being able to respond appropriately.
Spotify analyses the music you listen to and then uses machine learning techniques to offer personalised music suggestions with Discover Weekly. Netflix does the same for films and TV shows. Facebook’s facial recognition system uses machine learning and after tagging your friend’s faces just a few times is apparently 98% accurate in picking them out of new photos – the human benchmark is 97.5%.
Voice-activated assistants like Alexa are also powered by machine learning, getting more and more sophisticated as they are exposed to more and more data. In 2018 a number of people reported hearing unprompted ‘creepy laughter’ from their Alexa, and some people, jokingly seized on it as possible evidence that their Alexa was becoming, not just more intelligent but self-aware. Amazon themselves laughed it off as a response to a mistaken voice command rather than their AI gaining consciousness.
There’s no doubt that one day, possibly not too far away, AI empowered by machine learning, will become smarter than humans. Whether it could ever become what we think of as sentient – and perhaps decide that humans should be kept as pets or else disposed of altogether as some people suggest – we just don’t know. But let’s finish on a note of optimism. Google’s Director of Engineering, Ray Kurzweil, believes that not only will machines surpass us in intelligence by 2045, but that it will prove a boost to humans who will themselves become empowered by technology within them. Let’s hope he’s right – and in the meantime, perhaps start making friends with your Alexa just in case.