By Mark Richards.
Can we trust what we read any more? We have all heard the term ‘fake news:’ but as artificial intelligence becomes ever more powerful and the use of internet bots increases, will we be able to tell what is real and what is fake any more? Will there even be a place for humans on the internet?
In 2016 the UK voted to leave the European Union. No, stick with me, I am not going to write about Brexit.
But after that vote an online petition was launched, calling for a second referendum. The petition had well over 3m signatures – at the time, easily the biggest response there had been to any parliamentary petition. The problem was that the petition had been hijacked by internet bots being used to post fake signatures. Thousands of signatures appeared to come from countries like North Korea and the South Sandwich Islands – and yes, the South Sandwich Islands actually exist.
The simple fact was that whichever way you voted in the 2016 Referendum you were in no doubt that the other side was telling lies. They were peddling ‘fake news’ stories, whether it was ‘Project Fear’ predicting doom and gloom if we voted to Leave, or a bus trundling round the UK with ‘£350m a week for the NHS’ on the side. Or, as we have seen, an artificially inflated petition.
We are now – nearly three years on – seeing another petition, this time to revoke Article 50, under which the UK will leave the EU. This time the petition has passed 5m signatures – but once again there are plenty of accusations that thousands of the signatures are not genuine: that once again, the bots have been at work.
What is a ‘bot?’
A bot – otherwise known as a web robot or internet bot – is a software application that runs automated tasks over the internet. Typically a bot will perform a simple, repetitive task but at many times the speed a human could do it. There are already estimates that 50% of the current web traffic is generated by bots.
So will the internet eventually be a place without humans?
I have stated two or three times in these columns that I fully expect to be replaced by a bot within five years: that articles like this will quite easily be written with the aid of artificial intelligence.
But we can go one step further? Could the internet soon become a place without humans?
Right now Google allows us to find anything we want and Amazon allows us to buy anything we want. And depending on how far down the rabbit hole you want to go you can pretty much find or read any material you want.
But increasingly, the big companies are using AI to ‘help me organise the internet.’ Google is helping me organise my search, even – sometimes – gently suggesting that I have been searching for the wrong thing. Facebook ‘helps me organise’ what stories I will see.
These trends are only going to increase as more companies use bots powered by AI. We are already chatting to bots, relying on them for customer service and allowing them to influence our news.
Bots – hiding behind human names – are starting to analyse job or loan applications and politely turn us down if we do not match the profile they are looking for. And then ‘Lauren’ or ‘Joe’ will send us a recommendation on how we can improve our miserable CV or credit rating.
It is a short step from this to tech-savvy humans using bots to engage with other bots. Why not have a bot make the loan application for you? Why not have the bot make a hundred loan applications? Or apply for a hundred jobs?
The problem is – of course – where does it end? If we increasingly delegate tasks to bots who do those tasks by engaging with other bots, where does that leave humans? Could bots start talking about things we do not want them to talk about? Discussing our personal information with other bots? Could they even start talking in ways we do not understand?
It has already happened. In 2016 Microsoft had to shut down its chatbot, Tay, when it began to post offensive tweets. In 2017 Facebook shut down a project in which two AI bots seemed to be talking to each other in a language only they understood.
Whatever language bots talk in, one thing is undeniable: they are already being used to filter the news we read and the opinions we are exposed to – leading to more and more charges of fake news.
What is ‘fake news?’
Simply put it is stories that contain deliberately wrong information or hoaxes, usually spread via social media or online news outlets. Some of it is easy to spot – rarely a day goes by without a celebrity ‘dying’ on my news feed, or being found in a hostel for down-and-outs, or achieving a miracle weight-loss in a few weeks thanks to the latest wonder slimming pill. Websites need to attract advertising revenue and these clickbait stories work.
But we are seeing increasingly subtle fake news stories moving into mainstream media. It is becoming increasingly hard to tell truth from fiction – and governments are seizing the chance to control the news agenda.
Offend Vladimir Putin at your peril
In what is clearly a sign of the times – and something which must have several Western politicians salivating at the prospect – Russian President Vladimir Putin has signed a new law which will allow the punishment of online media and individuals for spreading information, deemed to be ‘fake news’ which ‘disrespects the state.’
That includes anything prosecutors consider to be insulting to government officials, with the new law declaring that ‘unreliable information’ undermines social order. It is difficult to escape the feeling that the definition of ‘unreliable information’ is ‘anything we disagree with.’ And the defence, ‘It wasn’t me, Mr Putin, it was my bot’ is not going to do you any good…
Couple that with China’s social credit rating – that we have written about previously – and you have two of the three biggest powers in the world making no secret of their aim to increasingly control citizens’ lives.
Meanwhile across the border…
Can we counter fake news and protect people against it? Across the border in Ukraine students are being given lessons in how to spot fake news stories, with reports that students who have received this literacy training are 18% better at identifying such stories. So education might be the answer but obviously, it depends on who is in control of the curriculum.
So where does that leave you and me? It leaves us in a position where we are increasingly going to talk to AI bots, not people. It increasingly means that those bots will filter the news and information we receive – and it increasingly raises the possibility that if we go looking for other information or entertainment, the government will want to know. After all, seven days from now you will need government approval before you can visit some of your favourite online sites…