We have all heard that the era of automation is on the horizon. We see the testing of automated vehicles. We hear the warnings of huge losses of jobs in the future. We’ve seen the estimates of changes automation will cause within the world economy. But, for the most part, writers have felt safe. Writing and blogging require creativity. It’s the foundation of the human skill set that is the most difficult to replicate using artificial intelligence.
But that doesn’t mean that writers should be without worry. There are specific types of writing where AI will replace writers almost entirely, and it is possible to imagine a future where the majority of written content we consume online is created by bots. In fact, we’ve examples of AI-generated stories being used in the press already. As time goes on, these types of solutions will find more applications. Anything that is dry and data-based can be automated, but companies have struggled to emulate true creative writing.
To understand what type of writing is most likely to be automated, we should look at what traits lend themselves well to automation.
What Positions Are Most Likely to Be Automated?
National Public Radio provides a free calculator that evaluates jobs and assigns a score that represents the likelihood that it will someday be automated. According to NPR’s automation estimation tool, there are four basic questions that you can ask about your profession that determine the likelihood that it will someday be automated:
- Do you need to come up with clever solutions?
- Are you required to personally help others?
- Does your job require you to squeeze into small spaces?
- Does your job require negotiation?
According to their tool, the likelihood of writing tasks being automated depends heavily on the type of writing. For instance, technical writers use short, structured sentences and use data to convey information. They produce product information, reports and internal documents. These tasks lend themselves well to automation. According to the NPR tool, technical writers have an 88.8% chance of having their jobs automated. The Associated Press’ use of automated stock market reports is a good example of the type of writing job that will be squeezed out by artificial intelligence in the future. However, other writing positions such as authors and creative writers have as low as a 3.0% of being automated.
Major Hurdles for AI Writing
Artificial intelligence has started to creep into every facet of business, but AI for writing has fallen a bit behind the curve for numerous reasons.
There are some major hurdles when it comes to writing automated content that make it a more difficult. Even today, straight-forward applications such as language translation are still iffy at best. Google Translate might be solid enough to convey basic messages, but it often provides translations that are full of errors.
There are numerous hurdles that have to be overcome before bot-generated content is truly accepted by the mainstream:
- Mastering analysis programmatically. Great writing is full of creative analysis. So while automated content might be able to put together basic factual articles, there is still a huge jump needed before AI is accurately analyzing data and conveying specific conclusions. Bots may be able to pull and report from a dataset, but will AI be able to draw realistic conclusions from that data? We have yet to see that kind of true artificial intelligence as of yet.
- Appealing to the reader’s emotions. People enjoy reading things that they connect with. This creates a question of whether AI will ever truly be able to connect with readers in the same way a person could. Being able to appeal to the reader’s emotions will be tough to master. Yes, content bots will have access to hordes of data that can provide insight on word usage. They will be able to analyze pieces and make recommendations. But, being able to put together an entire article that tugs at the heart strings is a tough thing to code.
- Making the content enjoyable. While AI might be able to master the reporting of facts, making the content enjoyable for readers has been tough. In certain types of writing, and specifically blogging, creativity is required to keep the reader engaged with the content. Will AI writers be able to conjure pop culture references that make the audience laugh? Will it draw comparisons to relatable scenarios? Until bots can produce blogs that intrigue readers, there will be a place for human-written blogs.
Examples of Bot-Generated Content Solutions
Despite major hurdles, there is no doubt that there is a place for bot-generated content in the future. The rollout of true AI-generated content will be slow, but it will come.
In fact, there are already several startups that offer solutions that are used to produce content by major companies. As of now, the focus of these solutions is on less-creative applications, but the foundation is being laid for bot bloggers of the future. Some of the major players in the AI-content generation scene that you should be familiar with include:
Automated Insights – AI-generated News Stories
Automated Insights is perhaps the most popular example of AI-generated content today. Their software aims to create natural language paired with data. They have already partnered with the Associated Press to deliver stock market reports and minor league baseball coverage. With this partnership, the AP has estimated that they will be able to deliver ten times as many stock market reports each quarter. This is a great example of how early implementations of AI-generated content might be deployed in the coming years. Simple data-based content that doesn’t require a lot of creative input to convey important points will be the first to find a foothold.
Persado – AI-generated Copywriting
Persado is another excellent example of AI-generated content. The startup focuses on marketing, copywriting, and conversion optimization. Their software uses data to deliver small snippets of content like product descriptions and short-form copy, then optimizes that copy to improve conversion rates. They claim an average click-through lift of 68.4%, and boast a total of 16 Fortune 500 clients. They recently raised $30 million in funding from institutions like Goldman Sachs. These kind of applications seem perfect for bot-generated content. The writing is mostly short-form and focused on improving metrics, giving the AI a clear goal for each piece. These smaller content snippets will be much easier to automate and are likely to be among the first to find a foothold in AI content generation.
CortX – AI-generated Web Content
CortX is a startup in the AI web content space. They are the umbrella company for brands like WordAI and Article Forge, which focus on creating longtail SEO content that sounds natural. These services are another great example of the future of AI-generated content, focusing more on simple, fact-based content.
The Future of Blogging – Where is AI Most Likely to Fit In?
So, are bots the future of blogging and journalism?
Probably not, at least for now. But AI will play a critical role in the development of online content in general. There are many hurdles that AI-generated content must overcome before companies are willing to trust them with long form content. However, there are a few ways in which artificial intelligence will help bloggers and content creators as a supplement to the creative process.
Short, fact-based Content
Before we see full-blown AI storytelling solutions, we’ll probably see a rapid increase in the amount of short, fact-based content that is generated using bots. Automated Insight’s recent deal with the Associated Press to produce robojournalism is a good example of the type of content that is ripe for automation. As these solutions improve, more advanced analysis could be possible.
Copywriting and Split-Testing
Artificial intelligence is perfect for marketing applications. The ability to have AI analyze your messaging, recommend potential variations, test against a control, and optimize for conversion is a natural progression for the industry. Automating short marketing messages like headlines and product descriptions is much easier than trying to automate the writing and editing of a 1,000-word blog article.
Creating great longform content requires that writers get their hands dirty. They may spend hours sifting through endless Excel documents to find interesting tidbits of data. But, sifting through and pulling from large datasets is where AI-generated content will shine. There are already solutions on the market that will allow you to pull interesting bits of data from a huge dataset with the click of a button. Connecting that data to a narrative is where humans excel. As time passes, we may see specialized tools for specific industries.
Should Writers Be Worried?
Most writing jobs are extremely competitive. The AI generated content of the future is only going to reduce the number of writing opportunities. Fortunately for writers, we are still quite a way from bots being able to replace a human touch when it comes to creative writing.
Writers should view early AI content generation as a tool in their arsenal, rather than competition. There may come a day when AI is able to supplant experienced writers, but there are other industries at much greater risk.