AI is already used in some journalism – such as at Associated Press and Thomson Reuters. Is there a limit to the type of content AI can deliver?
Tom Harvey, Account Executive, AprilSix Proof –
As AI continues to advance and develop, it isn’t out of the realms of possibility that it could one day take on the human journalistic roles that we wouldn’t have thought possible even a few years ago.
For example, tech experts are already predicting that in 2017, new AI personalities which have unique voice tones that align with core brands will be debuted through voice platforms. If we can implement AI that functions with a personality, that means that rather than just reporting factual news, technology will be able to offer opinions and insight into different topics, essentially making the human element of journalism automated.
What if the opinion section of The Guardian could offer the same diverse views and articles, but didn’t require a person to write them? Why spend money hiring a journalist to write these pieces when you can have a machine do it at a cheaper rate?
Danny Mitchell, Senior Account Executive, AprilSix Proof –
Correct, AI, or “robot journalism” is already being used to cover the most basic of stories – a short piece of copy about an earthquake or a report on a company’s earnings report. However, this is very basic reporting, which can be achieved by the “software aided reporting” using an algorithm to analyse sets of data, and then churning out a story by utilising pre-saved key words.
But this software does not look beyond the data and does not understand what truly makes a story. Because of this the stories that were written by robots (I refuse to call them journalists) are cold, bland pieces of coverage – as discussed in this interesting piece on Al Jazeera. Vitally they are lacking the most important element to any story, the human angle. It could happily churn out that a 6.8 magnitude earthquake has hit a city all day long but that’s all the story will contain – data. It won’t be able to tell you the stories from the survivors, the first responders, and will not be able to describe the sheer devastation on the ground. In essence, the stories are meaningless without the human angle.
Robots, of course, cannot investigate. They will not be able to build contacts, they will not be able to get on the ground and interview, and they will not be able to use their eyes to analyse the world around them. I was once told that any good journalist should be able to walk down any high street and come back with 10 story ideas either by speaking to residents or just by using their eyes i.e “Why is that building boarded up?” “Why do gangs gather in this one place?” “Why is there so much rubbish?” And this, in my opinion, is something any form of AI will never be able to do.
Do you think journalists are worried about losing their jobs? Or is it an opportunity to diversify and do more in-depth investigative journalism?
I think there will certainly be an air of nervousness for journalists going forward, especially as AI develops and becomes further implemented into our lives.
With regards to investigative journalism, what’s to say that technology won’t be able to do this? It was recently reported (I believe by a ‘human’) that CIA-linked firm CosmiQ is working with Amazon to develop a satellite that uses artificial intelligence to analyse detailed pictures from space. Imagine if this technology could be harnessed by the media, what would be the need for investigative journalists if images of the whole world can be taken, analysed and reported on from space?
Also, if AI has the capability to survey the whole world, what’s to say we couldn’t reach a point where it can hack phones or monitor people through webcams? AI journalists could offer a far more in-depth form of investigative journalism than their human counterparts.
Journalists are always worried about their jobs. Redundancy is a constant threat when you’re in journalism. But that is because the industry is changing not because of the rise of the robots. We are not seeing the beginning of some weird, dystopian film, “Terminator 6: The Robots Write Back.” The Press Association has been embracing automated journalism but only for the most basic of stories, and its own Editor-In-Chief has come out and said that it will not be used to replace journalists.
What this type of AI brings is an opportunity. It can save journalists time by analysing large sets of data or sending out alerts about an earthquake or a final score of a football match, allowing journalists to press-on with more in-depth, investigative reporting.
This is of course, depends on the software aided reporting being used correctly. I do fear that a publisher, who are often businessmen and not journalists, may only see the bottom line – see that they need less staff and can get a story out quickly – and push ahead with replacing journalists. If that did happen I think they would realise the mistakes they had made very quickly.
Can AI be a tool for journalists in developing stories? If so, how?
I read an interesting article recently from Kurt Barling, Professor of Journalism at Middlesex University, discussing how journalists need to harness artificial intelligence in order to tackle the rise of fake news.
Professor Barling argues that AI could be used to sample a range of versions of a story from various validated sources to create a data set, where algorithms can then be used to strip out bias and reconstruct the core, corroborated facts of any given event.
Now, while this isn’t exactly demonstrating how journalists can use AI to develop stories, what it does show is how Artificial intelligence can be used to make sure people are receiving factual news and not being hit with counterfeit websites and articles.
However, if AI can do this as well as write and report the news, then what is the need for journalists? This technology can help develop stories and news, but if it can produce the content as well, it renders journalists irrelevant.
AI is perfect for analysing large sets of data which humans would either find impossible or would simply not have the time to do. This opens doors for journalists to investigate new areas which may not have been possible in the past. It can also assist with short alerts, or push notifications, about breaking news – like an earthquake hitting a city.
Are there tasks an AI journalist can do that are beyond a human?
I wouldn’t say so as of yet, but I think in the future there certainly will be. I refer back to my earlier example of CosmiQ and Amazon developing a satellite with AI capabilities. If this technology can be applied in a journalistic landscape, the capabilities of AI journalists will be limitless.
As mentioned above, AI would be able to sift through large amounts of data and spot patterns that it would be impossible for a human to do.
What could the legal ramifications of AI journalists be? For example, can a robot be sued for libel?
There is always going to be ethical issues when AI comes into the equation, there has to be, we’re talking about self-thinking machines and we’ve all seen how that turned out in The Terminator.
I’m not sure how you would go about suing a robot for libel, but we could develop AI that understands moral and legal boundaries. If you could program AI journalists to have a detailed understanding of media law, such as what and what isn’t in the public interest, there would be no need for libel cases anymore as these machines will be incapable of committing acts that are deemed illegal in journalistic terms.
That’s the million-pound question. It has been suggested that it would, in fact, be the writer of the algorithm who would be responsible. But who knows, I’m sure it would rumble on through the courts for years. That is another reason, I believe, that this software will only ever be used for the most basic of stories.
Will an AI journalist ever win a Pulitzer Prize?
Why not? A robot has beaten the world record for solving a Rubik’s cube, can beat a human at table football and can outperform human surgeons, so maybe writing the next great novel or breaking an award-winning great story isn’t next?
No. Have you ever read that awful Friends script written by AI? If not, please do it’s awful. If gobbledygook likes this ever stood a chance of winning a Pulitzer then every journalist in the world may as well throw their notepads in the bin and go to the pub. Currently, AI is not sophisticated enough to write in-depth narrative reports and I am not sure it ever will be. The fact of the matter is that the very best journalism comes from journalists investigating, building contacts, having an eye for a story, getting on the ground to interview people and draw out the truth, and being tenacious in their pursuit of the story. These are things only a human will ever be able to do. To put it bluntly, AI would never have been able to uncover the likes of the MPs expenses scandal, Watergate, the phone hacking scandal, and much more.
As a PR professional, how would you pitch an AI journalist?
A good question. Unfortunately, I haven’t learned to speak in binary yet but I guess it would work along the same lines as pitching a human journalist. If AI is to replace humans in the press as I suggest, it is going to need the capability to understand what will make a good story and appeal to readers. That means from a PR perspective finding the relevant media angle from your clients and demonstrating this in a clear, concise way will still be the best approach to securing a story.
Much like Tom, I can’t speak in binary, so I won’t be pitching to an AI ‘journalist’ any time soon. AprilSix Proof could look into hiring R2D2 but I’ve heard he’s busy with Star Wars VIII. In all seriousness, I can’t see this scenario ever happening.