Artificial intelligence: good news and bad news.
That’s the title of a fascinating introduction to AI by Fabio Lalli – a friend of mine and one of the pioneers of the Italian and European digital revolution.
Fabio wrote the article – which weighs up the positive and negative impacts of AI – drawing on his experience running courses for creatives, professionals and academics and weeks of experimentation, including generating hundreds of images and texts and integration tests with third-party scripts and tools. He comes to two conclusions, or, as the title states, some good news and some bad news:
Good news: AI will not replace us as professionals or businesses.
Bad news: a person or business using artificial intelligence will.
The rapid development of generative artificial intelligence: from UGC to AIGC
But let’s take things one step at a time. What has prompted Fabio and many other experts on the crossover of technology, communication and marketing to consider this issue? It’s all down to the latest buzzword within (and outside) the sector: generative artificial intelligence.
This is where AI algorithms create new content from existing text, audio files and images. The generative AI process allows machines to deduce the underlying pattern or template of the input material and then generate a similar, plausible version or copy based on this information.
One of the key factors that fuelled marketing in digital markets was undoubtedly the ability they gave people to create and share user-generated content (UGC). The various platforms and social networks are continuously being fed with content, in a never-ending stream of shared personal experiences.
However, a new type of content is already emerging and spreading in parallel to UGC, generated by artificial intelligence: AI-generated content (AIGC).
Interesting AIGC examples and tools
It may surprise you to hear that machines have been able to create content for several years now, and are already widely used for that very purpose. Various technologies, tools and practices in this field are becoming part and parcel of brand communication and marketing. Take, for example, the content created for the Italian market by Mulino Bianco using Midjourney (AI software invented by the lab of the same name that produces images from textual descriptions) to advertise its pancakes. They are one of the first companies in the world to use this technology: you can see them here (Facebook) and here (Instagram)!
Midjourney has been one of the most talked-about tools in recent months, but there are many others that could become essential for your creative projects in 2023 and beyond. Here are five more that will certainly be of interest to any communication professional:
- cleanup.pictures, used to remove unwanted objects, people, text or defects from any image.
- soundraw.io, a music generator for creators that allows you to select your preferred type of music in terms of genre, instruments, mood, length, etc, and let AI write beautiful songs for you.
- looka.com, a tool for designing an effective logo, website or brand identity using nothing but the power of AI.
- copy.ai, to produce high-quality copywriting that can sell, persuade and convert clients.
- scriptbook.io, a startup founded in 2015 in Belgium, which has developed an experimental tool that uses AI and machine learning to analyse film scripts and predict their success or failure three times more reliably than human analysts.
There are, of course, many more, including those summarised in the visualisation below – a constant work in progress!
These tools are still being optimised, and should be assessed not so much (or at least not exclusively) in their current state, but rather in terms of the horizons they could open up in business communication and elsewhere.
According to the founder of ScriptBook Nadia Azermai, for example, Sony would have saved a great deal of money if it had used their algorithm instead of people when approving film pitches. The Scriptbook software was able to retrospectively identify 22 of the 32 Sony films that ended up being flops in that period (movies that passed traditional human-based screening, including experts, focus groups and market research).
The mission of the AI content creator Articoolo makes interesting reading:
Winning content marketing. Our technology creates unique content from scratch, simulating a human writer.
And the goal of ArticleForge is equally striking:
Get high quality content in one click.
These solutions simulate the work of a copywriter, but using a machine: machine-made content, in other words.
UGC and AIGC: which is better, word-of-mouth or word-of-machine?
Anyway, back to UGC and AIGC.
Are we sure the content of the future will necessarily be created solely by bots and machines?
Not necessarily… and not the content of the present, either 🙂
Chiara Longoni from Boston University and Luca Cian from the University of Virginia have been looking into just that: they recently carried out a research project involving 10 experiments on 3,000 people, published in the Journal of Marketing and also picked up by the Harvard Business Review. As an alternative to word-of-mouth – which today plays a decisive role in any customer journey – the two researchers talk about word-of-machine, meaning the situations where we prefer the advice generated by AI over that given by other people, or at least take it seriously into consideration.
- Generally, if our buying aims are utilitarian or concentrate on the functional characteristics of the thing we are purchasing (a dishwasher, for example) we trust the advice of machines.
- If, meanwhile, the decision is based on experiences, or hedonistic or sensory aspects come into play (wines or perfumes, for example…), the advice of AI is no longer enough, and has to be combined with a human touch. This has been implemented successfully in various fields, such as Stitch Fix, a personal styling service that integrates advice from other people and algorithms.
Synergies can also arise between UGC and AIGC: this is precisely what happened with Zo, Microsoft’s social AI bot, which in October 2018 reached Wattpad – an online community for writers dedicated to creating user-generated stories – and, after overcoming some qualms from human users, hosted the contest #WriteWithZo.
The logic was simple: Zo encouraged people to interact on Facebook, Twitter, Skype, GroupMe and Kik to get recommendations in the form of ideas for titles, characters and settings for different literary genres. The results were surprising: in just over six weeks, 150,000 stories were generated and uploaded to Wattpad by inspired writers from over 800 countries.
A long road ahead
It seems we have a great deal of change ahead of us. For a while now, global businesses like Associated Press, Yahoo!, Forbes and Reuters have been using algorithms and non-human intelligence to plumb big data and work out the most important topics being discussed and spread by online audiences, and so generate relevant content in real time. As noted by the online magazine Futurism, if you happen to read a CNET article with the generic sign-off ‘CNET Money Staff’, try clicking on the author’s name: in the biography, you’ll see: ‘This article was generated using automation technology’. Meanwhile, research at the MIT in Boston, one of the most famous and authoritative universities in the world, has been imagining profitable collaboration between human journalists and robo-reporters for years.
Back during the 2016 Rio de Janeiro Olympic Games, the content and information arena was populated by all the world’s top media outlets. Toutiao, a Chinese news platform with an audience that is 90% comprised of young people aged under 30, published 450 articles – 30-40 articles per day for 15 days of competitions – by unleashing Xiaomingbot, an AI robot that specialises in writing articles and sports content. The Washington Post then responded with Heliograf, a bot that can write sports articles and keep track of medallists’ progress in real time.
As users accessing this content, we don’t tend to notice it wasn’t written by humans. And the costs are tiny (efficiency) and the capacities are infinitely larger (effectiveness) compared to the alternative of journalists, copywriters and other communicators.
There’s a fine line between copywriting and data science… so it’s worth us all being prepared. As communication professionals, because we’ll need to be able to combine our skills with more high-tech approaches and ‘augmented’ marketing. And as business people, because increasingly we’ll have to assess designs and products that may not have been developed by humans. And because, as Harvard Business Review reminds us, algorithms need managers too.