Artificial Intelligence – 10 things you never knew!
By Paul Field – CEO @ Improved Apps
It is very easy to get swept along with the growing tide of news about artificial intelligence (AI) and start to think that the prospect of an AI that could out-think a human and reason independently and originally is a close prospect. However, we are still a long way from an AI that could in any sense be said to be sentient.
Nevertheless, there are clearly AIs already out there in the world around us performing far better that humans can: More than 75% of share trading is automatically handled by AIs, flights are kept safe from each other thanks to AIs, and even medical diagnosis has been helped by AIs. AIs can beat us at chess or Jeopardy and provide customer guidance on web sites.
All of the above though are examples of ‘weak’ or ‘narrow’ AIs (known as ANIs as opposed to the AGI that represents general intelligence). Deep Blue couldn’t play you at Monopoly, air traffic control couldn’t find your best route to drive home, and share trading systems couldn’t help you plan for your retirement. The AIs that have been created so far have been dedicated to one function or domain and have been able to use the superior number-crunching/alternative-evaluating capabilities of the computer to edge ahead of humans.
Estimates vary, but rarely put the advent of a genuine AGI before 2050. This still does not imply anything about genuine self-awareness or sentience – if indeed that can ever be achieved. We also like to confuse ourselves between what the AI might be housed in (a robot for instance) and the computer that is being housed. AI is ‘just’ a computer (or network of them) and – in comparison with humans – a very young and impressionable being.
Humans learn by making mistakes, by trial and error and an exploration of different possibilities. Having an AI learn by its mistakes when it is doing something important is not something particularly desirable.
So, rather than try and predict how ‘junior’ will grow up, here are a few nuggets of information that might be of interest:
In the 1960s, an average share of stock was held for 4 years. By 2000, as a result of automated trading, average ownership dropped to 8 months, and in 2008 it dropped even further to 2 months. Today the average share is held a scant 20 seconds and within a few months, it will drop to less than 10 seconds.
Flash Crash: It was 6 May 2010. In the UK it was general election day, in the US, Wall Street was gripped by mounting anxiety about the Greek debt crisis. The euro was falling against the dollar and the yen, but despite the turbulent start to the trading day, no one had expected the near 1,000-point dive in share prices: In a matter of minutes the Dow Jones index lost almost 9% of its value – in a sequences of events that quickly became known as “flash crash”. Hundreds of billions of dollars were wiped off the share prices of household name companies like Proctor & Gamble and General Electric. But the carnage, which took place at a speed never before witnessed, did not last long. The market rapidly regained its composure and eventually closed 3% lower. This whole process was driven by the algorithms controlling the trading systems.
Joshua D Brown – you probably have never heard of him – but he was the first person to die in a self-driving car accident. His Tesla drove into a lorry when on auto-pilot.
AIs can already write – but not really create. The following piece has appeared in the Los Angeles Times:
“A shallow magnitude 4.7 earthquake was reported Monday morning five miles [8km] from Westwood, California, according to the US Geological Survey. The temblor occurred at 6.25am Pacific time at a depth of 5.0 miles. According to the USGS, the epicenter was six miles from Beverly Hills, California, seven miles from Universal City, California, seven miles from Santa Monica, California, and 348 miles from Sacramento, California. In the past 10 days, there have been no earthquakes magnitude 3.0 and greater centered nearby. This information comes from the USGS Earthquake Notification Service and this post was created by an algorithm written by the author.”
Had one of those birthday cards that plays a tune when you open it? Did you drop it in the bin after your birthday? You’ve just thrown away more computer power that existed in the whole world before 1950.
Nautilus, the SGI Altix supercomputer is constantly fed with news articles dating back to 1945. What it does is read the mood of the article and location, which it then weaves into a spider web of relations. Connecting one of these threads, it then predicts the future. Nautilus predicted the Osama Bin Laden future hiding place with an accuracy of 125 miles and also predicted the 2010 Arab Spring.
Ask Cortana for a joke – I got “Why did the chicken go to the séance? To get to the other side.”
In early 2016 Microsoft released an intelligent chat bot called Tay: Tay was built to speak like a teen girl and released as an experiment to improve Microsoft’s automated customer service. Instead, “she” turned into a complete PR disaster – within hours of being unleashed on Twitter, the “innocent teen” bot was transformed into a fascist, misogynistic, racist, pornographic entity. Her tweets, including phrases like “Heil Hitler”, were disseminated widely as an example of why Twitter reflects the worst of humanity.
A computer program called Eugene Goostman, which simulates a 13-year-old Ukrainian boy, is the first that is claimed to have passed the Turing test at an event organised by the University of Reading. The test investigates whether people can detect if they are talking to machines or humans. The experiment is based on Alan Turing’s question-and-answer game ‘can machines think?’ The 65-year-old test is successfully passed if a computer is mistaken for a human more than 30% of the time during a series of five-minute keyboard conversations. (It is worth noting that this ‘win’ has been heavily contested). Eugene convinced 33% of the judges at the Royal Society in London that it was human.
In the years to come, an AI probably WILL take your job…
Paul Field – CEO @ Improved Apps