Most of us are familiar with the concept of autonomous cars and how artificial intelligence (AI) technology is likely to make human driving, along with all its inherent dangers, a redundant task. What is less well known is how the capabilities of AI are also growing rapidly in shaping the creation of the written content we consume in our everyday lives.
An increasing amount of what we read is being created with some machine input, and this trend is highly likely to continue. For example, a writer might run their content through a quality assurance (QA) tool, and depending on its sophistication, this tool could perform a variety of tasks, from checking spelling and grammar to assessing the content against a set of customized rules based on style and tone requirements. In other cases, if we’re looking at company reports or articles with detailed analytical content, it’s quite likely that natural language technology—a form of AI—was used; the AI will take structured data sets and convert them into narrative text that a human can easily understand.
How does natural language technology work?
Before we talk about AI in content creation, let’s look at the three components of natural language technology and how we might already be familiar with them in our everyday lives.
- Natural language processing (NLP) refers to interpreting and analyzing data that is in the form of natural languages, whether in speech or written forms. With the right programming, NLP can take source material in any language and interpret it in a highly standardized way. Examples of NLP in action include auto-correction tools on our smartphones and email spam detection.
- Natural language understanding (NLU) processes and derives meaning from language input. This might refer to the responses spoken during an automated telephone call, an automated reply to an email to a customer service department, or searching a huge volume of text in legal documents. The AI system can derive sentiment and context from text, drawing usable data from the content. An example is social media monitoring: a company can apply an algorithm to identify not only when their brand is mentioned, but in what context, and whether these mentions are associated with praise or criticism.
- Natural language generation (NLG) is the conversion of the data derived from NLP and NLU into written (or spoken) insights, analysis or factual content. This is delivered in natural, easily understood text or speech, and can be presented in multiple languages as needed.
So how is natural language technology currently being applied to content creation, and how is it likely to shape our world in the near future?
Books written by computers
In 2016, a team led by Hitoshi Matsubara, a computer scientist at Future University Hakodate, created an AI program that co-authored a short novel that made it through the first round of judging for the Hoshi Shinichi Literary Award. In this case, a human team created the plot and structure, and the computer pieced together the components of the story. The finished work was considered on par with that of established writers—an amazing feat and powerful evidence of the potential of NLG.
American author Hugh Howey has suggested that the success of the Japanese team could be a model for computers writing whole novels in the future; indeed, it won’t be a huge step for AI to absorb the remaining task of coming up with a plot, thus eliminating human involvement entirely. The idea of browsing through new releases in a bookstore and not knowing which books were written by “real,” human authors is no longer a far-fetched one.
Where is this new technology already applied?
Perhaps one of the most common uses of natural language technology is in writing product descriptions. By their very nature, these need to follow a consistent format in which a product’s key features and benefits are presented in a way that’s easy to understand and compare against alternatives. At Moravia, we’ve found that, providing that the data is sourced in a structured way, it can be processed and delivered to the user as text that consistently meets a bespoke set of rules; we can even ensure that the content follows our clients’ tone of voice and brand guidelines.
Not only is the activity of creating thousands of product descriptions converted from a costly and inconsistent process to a simple automated task that can be replicated in multiple languages, but any updates can be readily applied simply by modifying the source data, as long as it has maintained its connection to the NLG text.
Another example: we increasingly see companies using chatbots as a front-line customer service tool, allowing them to identify requests or problems and then either provide responses or escalate the issues to human employees. As these chatbots develop, inevitably they will be able to take on a growing number of queries, and the need for human customer service staff will decline; chances are that as a customer, you won’t even know if you’re dealing with man or machine. If your main experience with natural language technology is in the form of a robotic voice giving you a list of options when you call your utilities company, this demonstration of Google Duplex might surprise you.
Perhaps the most powerful use of natural language technology is in processing and analyzing large amounts of data, such as in the creation of company reports or legal documents, turning these into concisely written content with a structured narrative. This type of work traditionally takes a person many hours of gathering and sifting, followed by number-crunching and report-writing. Now such a project can be almost entirely automated, saving companies both time and money.
Will the copywriter be consigned to the scrapheap?
On the face of it, the future doesn’t look bright for the professional writer. Is technology really on the verge of killing the need for a skilled scribe? If you’re an established copywriter, there’s certainly no need to panic—at least not yet. Natural language technology will bring new opportunities for writers, and if you want to take advantage of them, it’s more important than ever to stay on top of these technological developments. A likely scenario might be that while you’re typing your draft article, an AI engine is simultaneously interpreting your words and delivering relevant data and search results onto your screen, automating the laborious research element of your task. As a writer you can then focus on creating compelling, original content that emotionally draws in and persuades your readers. That’s something that, at least for now, technology can’t do.
What’s next for content creation?
As natural language technology continues to impact content creation, particularly in a multilingual environment, the effects are likely to be profound. More and more of the material we read will be created without a writer’s direct input, and the quality of most machine-generated content will become impossible to distinguish from that of a seasoned copywriter. Natural language technology offers enormous opportunities for businesses to produce large volumes of content, and reaching a global audience is no exception.