eg. Facebook or Google Analytics
Writing Content Photo by John Schnobrich on Unsplash 1080

Yohei Guy - Tuesday 22nd January 2019


AI: The ultimate content creator?

The idea of Artificial Intelligence (AI) is nothing new having been around for well over 60 years. Although hardware deficiencies curtailed its development for much of that time, we are now at a point where advances in technology allow us to ask some interesting questions.

We know for a fact that AI can perform and accomplish mechanical tasks with ease like assimilating and analysing large datasets. This is already being applied across many aspects of digital marketing from chatbots to EDMs, SEO and web design.

But how does AI fare when it comes to processes considered more abstract like forms of creative/artistic expression? Is it possible for AI to create in an original way like human beings? Can that ability be utilised in some way to create or enhance content?

If you’re involved in Content Marketing for an agency/organisation or as an influencer, the answers to these questions will be of great interest indeed.

AI and visual art

AI has been used to create visual art since the 1970s when Professor Harold Cohen from University of California San Diego(UCSD) created AARON, a computer program that was able to create original art, much of which was featured in galleries all around the world.

Similar efforts have been replicated in more recent times too with OBVIOUS in France, a collective that uses Generative Adversarial Networks (GANs) which are machine learning algorithms to create images. These images have sold for thousands at auction, a sign that AI art is starting to be recognised and accepted on its own merit.

This has relevance because similar AI is currently being used for image recognition in areas like social media where companies can analyse posts featuring images to see if there are shared characteristics in successful posts. They are also using brand recognition software to identify where their branding is featuring online and in what contexts, again pulling out successful examples to inform strategy going forward.

AI and storytelling

Individuals and organisations have had a crack at storytelling as well with some fascinating results.

One of the most well-known examples is the short film sci-fi script that was written by ‘Jetson’. This AI machine was fed hundreds of sci-fi screenplays and asked to create one of its own. ‘Sunspring’ was the resulting script, which was then produced and shared on Youtube.

The film itself isn’t the most coherent to follow with its cryptic and often bizarre use of language. However, the end result is undeniably captivating and unique and its 1 million views on Youtube demonstrates that it did successfully capture the interest of the public.

Although a feature film project hasn’t quite hit the screens yet, a Kickstarter campaign has been launched for one (co-written by AI) and one wouldn’t be surprised if a fully fledged film isn’t far off the horizon.

Establishing AI’s ability to tell stories is important as effective storytelling is the key pillar of compelling content. Whilst these examples produced slightly more surreal narratives the output was nevertheless intriguing. With further development, it’s not hard to imagine companies being able to input a number of motifs/themes aligned to their brand with a storyline produced in a specified format.

AI tools for content creators

As many of you will know, there are already many AI tools already in use by agencies and individuals in ways to shape, guide and deliver content more effectively.

The use of chatbots, as previously mentioned and the analysis of their conversations with tools like Dashbot and Bot Analytics can lead to some great ideas for relevant content. Tools like Acrolinx can also help writers with tone of voice and plugins like WordLift help optimise content in real time.

There have been other instances where AI has enhanced existing content as well with an interesting example being the 20th Century Fox film ‘Morgan’ made in 2016. This involved IBM Research using AI to identify moments in the film suitable for the movie trailer which was then cut together. The result was extremely convincing with the main benefit being it cut a process that usually takes weeks down to less than 24 hours.

AI as content creators

Enhancing is one thing, but full content creation is another. One area in which AI as a content creator currently exists is in journalism. The Associated Press, Yahoo and the Washington Post have all used automated journalism, quite prolifically, to cover sports matches, elections as well as reporting on finance.

The common trait for all of these topics is that they are statistical, data-backed pieces from which templates can be built and a piece created. Similar software has been developed and used to assemble reports for financial institutions and law firms, saving valuable time and costs during an otherwise protracted process.

With regards to more creative genres and forms of content, AI content creation tools and platforms have started to appear but are still largely evolving and offer different types of services.

  • Articoolo, which is in beta, creates content from scratch when given a topic and word length. It does this by finding sources of related content and then basically rewriting it using its Natural Language Processing engine.
  • AI Writer, in alpha phase, is more of a research tool that cites a variety of sources about a topic in order to help stimulate ideas for content.
  • Word AI, in its 4th version, is a rewriting tool that specialises in rewriting content in a way that is readable to people and recognised by Google as unique copy.
  • Article Forge uses deep understanding algorithms, reading millions of articles to create unique and original ones of its own.

How is AI impacting on brands and influencers?

AI is becoming more commonplace that’s for sure and the most forward-thinking brands are starting to explore the potential for AI to supplement the outstanding work already being done by their in-house content teams. For influencers, AI can be a little trickier. At the moment, cost and time can be prohibitive to really digging deep into the use of AI. The time it takes to set up the tools and to maximise their potential is often prohibitive for influencers although there are exceptions. Over the next couple of years, we would expect more automation and better platforms which make it easier for smaller businesses and influencers (as well as the bigger brands) to really start exploring the full potential of AI.

Conclusion

So where does this leave us right now as we approach the end of 2018? Well, whilst the platforms we have discussed all boast working examples and include testimonials on their sites, further searching online can reveal somewhat mixed (often hilarious) results. One can’t be too critical as the use of AI in this way is still evolving and many of these platforms are still in some sort of testing phase.

The real question is how much longer will it be before AI supercedes our ability to write? Well according to researchers at Cornell, they’ll be writing high-school essays by 2026 and writing best-sellers by 2049. For the most part, that means we’ll continue to write great content for our clients (and ourselves) for the time being, using AI tools to enhance the good work we do.

Leave a Reply

Your email address will not be published. Required fields are marked *