The machines amongst us
Around 1770, inventor Pierre Jaquet-Droz invented ‘The Writer’, a manikin in the form of a small boy who could dip his quill into an inkwell and accurately write on paper, his eyes following the words. Tucked within his body was an inscribed plate on which the alphabet could be rearranged to create different messages. As early as 1863, Samuel Butler theorised that at some point in time, machines would become so advanced that they would self-replicate and take over the world. In 1920, a play by Karel Capek called R.U.R. depicted slave androids rebelling, forever changing the Czech word for ‘serf’ into a futuristic term, ‘robot’. More recently, movies such as Bladerunner and Ex Machina and TV shows such as Black Mirror and Westworld have continued our fascination with the possibilities of AI. But in 2020, we’re at a point where it’s not just content about AI that continues to stoke our endless fascination about possibility of artificial minds; we are in that future and content created by AI is now here, amongst us. We can finally start to understand what they might have to say and how they might find ways to say it.
Relationship mapping will ultimately let us see unrelated themes which appeal across our audiences.
But first, some stats
Kai-Fu Lee, former president of Google China and author of AI Superpowers, thinks that around half of all existing jobs will be technically and economically doable by AI and automation over the next fifteen years. In 2016, Gartner predicted that chatbots would power 85% of all customer service interactions by 2020. 72% of businesses in the technology, media and telecommunications industry expect AI to have a significant impact on product offerings in the next five years and 75% of what users watch on Netflix come from its AI driven recommendations.
Translation services will help us to matter to those who think and dream in other languages.
Where are we now?
We’re already encountering content via machine learning and AI in our daily lives. Natural language processing is driving chatbots, our smart speakers and our spam; it’s a key driver of content algorithms, enabling social networks such as Facebook, Twitter and Instagram to customise our news feeds and the ads we see; and it’s creating a highly personalised curated universe. Netflix apparently created over 2,000 different assets to promote series three of Stranger Things, so, if you like to watch romcoms, you’d see images of the romantic leads but if you preferred horror, you’d see gorier tiles. Netflix, incidentally, claim they save $1bn a year by using AI.
From an editorial content perspective, companies such as the Associated Press, Yahoo and Fox have been using AI to generate content for a while. PA produces 30,000 local news stories a month using AI, a service it now sells to other content outlets under the brand RADAR. Roughly a third of content generated by Bloomberg News is created by some form of automation.
AI is here then. How can it help us, in the short term, improve our content output?
Each AI will need a personality to differentiate themselves from others in the market, so tone of voice will be crucial.
1. It will help alleviate some of the most boring publishing tasks
One of the reasons Bloomberg introduced its robot reporter, Cyborg, was to release its writers from the mundanity of repetitive research; sifting through endless financial reports and churning out stories focused on nothing but detailed facts and figures. Not only is AI an uncomplaining staffer, but it is fast and this advance has increased our ability to publish immediately.
Another task we’ll see disappear from our ‘to do’ lists will be in how we sort our content. All of that manual reviewing, tagging and filing will belong to AI. Natural language processing is finally at the stage where it can read our content, understand key themes and identify authors, read any copyright restrictions and assign multiple and complex tags.
The added benefit to such sophisticated sorting means that AI will be able to identify and surface content to us. For example, when we want to find images related to a piece we have just written, AI systems will not only be able to recommend related photos (and crop them), but will recommend assets based on past performance data for a particular audience based on the next action we want a reader to take. AI will also find related articles beyond simple headlines and tags, ensure we select the best formats, audit existing content, help us respond to real time customer needs and ultimately, free us up to be more amazing at creating.
2. Incredible insights will increase our creativity
One of the most advanced sectors in artificial intelligence is, well, intelligence. As data is the driver of machine learning, it has the obvious edge on development. AI is incredible at analysing huge swathes of data and identifying patterns and behaviours which would take analysts much longer to establish.
Data tools from SEO, site analytics and social have been helping us create relevant content for years, but AI is now helping us to understand what might happen next. Predictive intelligence will allow us to highly personalise what content our audiences see, by establishing their needs and interests via a points system or ascertaining a customer’s next likely decision based on other customer data.
AI will also start to identify possible relationships between our audiences and their wider interests to help us find new angles or ways to approach our work. Relationship mapping will ultimately let us see unrelated themes which appeal across our audiences and bring them together via our clever editorial methods.
3. It will help us create amazing personalised experiences
Personalisation is made up of three key layers: ‘Behavioural’ which is what readers look at, interact with, like and share; ‘Preferences’ where readers tell us about themselves; and finally, ‘Contextual’ which is whatever third party data we might overlay such as location or the weather. By bringing all of this data together and using previous customer data, we can start to understand and serve our audiences at a one to one level.
As digital asset management also develops AI, we’ll be able to identify an individual customer and serve a piece of content to them in a much more personalised way so that the image they see reflects them or the headline they read appeals directly to their persona. It’s like language translation but tailored to customer personality.
4. It will help us reach new readers
On the subject of translation, auto-translation is probably going to be one of the first sectors to see the replacement of humans by machines. This is primarily because AI learns a language in exactly the same way as children – by seeking patterns.
For content producers, it’s a huge opportunity as our potential audiences expand across the world. A recent study indicated that English-only content is a real barrier to access for the next billion, the newest wave of online consumers who hail from Asia. Translation services will help us to matter to those who think and dream in other languages.
On top of this, the much more highly developed targeting engines will help us put our content in front of those who may not have heard of us, but will inevitably respond positively to our content, because we are tightly matched with them.
5. It will empower because we’ll need to destroy the echo chamber
It’s hard to tell at this stage whether this is wishful thinking, but AI could either take us down the route where all we see is predefined and divisive or it is going to broaden our senses by creating bigger universes for us to explore or focusing on ways to create more two-sided views of subjects.
It’s likely that the role of curated content will flourish as customers, living in a post-truth world and increasingly aware of algorithmic confounding (when consumer options become increasingly narrower, and user choices can be restricted to increasingly extreme content), demand human expertise to help them find what they want.
Machine learning and AI both rely on information and data to help them function, but that data and information has to originate somewhere. Original, creative thinking is not currently something we’re seeing from AI; you only have to follow Janelle Shane’s incredible blog (AI Weirdness) to understand that what AI sees as creativity, the human world sees as plain strange: her AI recipes include Crock Pot Cold Water (“by my reckoning this produces a very gluey, greasy mess, tasting strongly of vanilla and baking powder), Chocolate Chicken Chicken Cake and Completely Meat Circle.
AI … will recommend assets based on past performance data for a particular audience based on the next action we want a reader to take.
AI, are we human or are we dancer?
In the short term, I believe that AI presents a huge opportunity for us. Each AI will need a personality to differentiate themselves from others in the market, so tone of voice will be crucial. Machine learning needs creative and original thinking, an origin story to tell. The data available to us could see our creative minds grow bigger than ever. Not having to do the mundane tasks can free us to be only creative.
However, I admit, like the Luddites past, that I’m worried about AI in the long-term future. I’m worried about people’s jobs, about displacement, about the scores of unskilled who may no longer be able to work. I’m not one of the eternal optimists who believes that by replacing the repetitive, we all have more time to read and learn and think.
We need to factor in and focus on how longer term, we ensure AI is our friend and not our foe. How it will only serve to protect our creativity, broaden our universes and push our boundaries. As an industry, it’s time to work together, to create codes of conduct which protect us from joblessness and save us from the echo chamber and the post-truth world and to ensure our livelihoods are protected.
Because we’re human and dancer.
It’s time to work together, to create codes of conduct which protect us from joblessness and save us from the echo chamber.
This article was first published in InPublishing magazine. If you would like to be added to the free mailing list, please register here.