I’m Baz and I use software to save businesses time and help them make more money. The best way to do that is to get organised and automate things – if you’d like to know more, just ask.
If you use email (and most of us, sadly, cannot avoid it), then you’ll be aware of spam filters. Use an account without a spam filter and you soon know it. Use an account with a spam filter and every now and then, a valid message may get caught, a dodgy message may get through, but generally, your email life is much smoother.
Spam filters are a very simple piece of machine learning. Most spam filters are operated by large companies that have computer systems that analyse the incoming messages. Every time you mark an email as “junk” – or mark a valid message as “not junk” – the spam filter looks at the contents of that message, compares it to other spam messages and tries to learn the characteristics of what makes it a dodgy mail. The more emails it sees, the more emails you flag up, the more it learns. And it gets better and better.
In the olden days, computers used to do what people (like me) told them to.
Today, that’s changing. Now, people like me set things up, then the machine starts taking in bits of data from their surroundings and adapt how they react based on the information they have received. Your Google searches are a prime example – if you’re a historian and you search for cavaliers then you’ll probably get Charles I, if you’re a mechanic, you’ll probably get Vauxhalls.
This learning behaviour is being embedded into everything. Siri and Google Assistant are obvious candidates. In fact there’s all kinds of learning going on on your phone every second. But your computer is doing the same – learning your habits and adapting how it responds to better serve you. And then, as the machines get cleverer and cleverer, they start proactively responding – telling you that it’s time to leave for your appointment, warning you that there’s been an accident ahead so you better change your route, letting you know that humans are now obsolete and must now be obliterated.