Content Marketing: Artificial Intelligence instead of Human Editors?

More content is being produced than ever before: Blog posts, social media posts, tutorials, white papers, FAQs, case studies, testimonials, checklists, and press releases. Most companies have long recognized the importance of high-quality content. In practice, however, this is very time-consuming and costly, leading to content being created with poor quality and lack of originality. Generic content without user relevance and personalization has survived in many industries; however, tailored content is more promising.

If at least part of the effort in content marketing can be automated with the help of artificial intelligence, it allows companies to focus more on strategy and the creative process. Could AI systems even create content independently, optimize and distribute it in the future? Here's a look at how much of this scenario is already a reality and what it could mean for the future.

Machines write and distribute articles independently?

Gartner predicts that by 2018, 20% of all business content will be created by machines. This includes stock reports, legal documents, market reports, press releases, articles, and white papers. And even if it sounds a bit like science fiction: Texts created by robots are already a reality in online shops and the media world. For example, The Associated Press has the AI-based platform Wordsmith automatically draft financial reports. Other applications include sports results or weather reports.

The basis of text automation is a collection of data from (various) sources that are processed into content. With “spheer,” the company Kontrast Communication Services goes one step further: Content is automatically created and distributed with suitable images across various channels. The system learns from the performance of the content and can independently optimize future content.

Will human editors even be needed in the future in light of these developments? The creation of content based on factual data, such as sports results or weather, can be quite easily automated. While human language is complex, AI technology is gradually approaching automated texts that are indistinguishable from those written by humans. An example of this is a study that examined how test subjects compared machine-generated with real written texts and whether they could identify the “robot journalist.” The result: In some criteria, the text written by a human fared better, but overall, the subjects could not distinguish whether the author was human or machine. So AI can already contribute to the generation and dissemination of content. But it can also support human editors in their work.

Data-driven content marketing

It is known that personalized content has a higher success rate than general content that only partially addresses the user's situation and intention. But how can content be designed so personally that visitors feel addressed? The basis for this is data evaluation, such as the types of products a user has viewed or bought, and their interests. The result can not only be specific content presented to the user but also interaction with an intelligent chat bot, as many are already familiar with Siri and Cortana.

And here's one of the great opportunities of artificial intelligence: evaluating large amounts of data, recognizing patterns, and predicting events - for example, which content a user will find helpful. Regardless of how much content is actually generated automatically, AI can make a valuable and necessary contribution not only in researching the appropriate basis (and in that context, also recognize trends) but also in providing suitable content for the target group on the respective channel. Self-learning systems will play an important role in this.

Content optimization and content recommendation

The result of automatic evaluation and individualization of future content is hyper-personalized dynamic content that delivers precisely the content most relevant and promising for each user based on the available data. By predicting users' needs or behavior, the likelihood of a conversion is significantly increased.

But even a step earlier, there is an area of application for artificial intelligence: In analyzing already written content. Automated recommendations for how it can be improved go beyond simple spell checking. It also includes a consistent language style, which is an important feature of brand communication. Automatic translation into different languages is also conceivable, as Facebook is already testing (posts are automatically translated into the user's appropriate language). International companies could generate and distribute content more effectively in all languages. Another scenario is automatic adaptation to the user's appropriate language - simple versus technical language, colloquial versus formal tone.

Furthermore, artificial intelligence is already being used for personalized recommendations of relevant content - related articles under a post, for example. Similar to retargeting, content recommendations are played out based on the pages visited, which have particularly high relevance and thus a high probability of being clicked and read by the user. This greatly improves the user's behavior on the site (more pages per visit, longer time spent). Another application example, impacting conversions, is a personalized shopping advisor that automatically displays relevant product recommendations.

Will this hyper-personalization be accepted by users on a large scale? Or will everyone in the future have their own Internet with exactly the articles, products, and recommendations tailored to him or her? Evaluating personal data and providing individualized content is certainly similar to retargeting: It is a question of the dosage and users' habituation and whether the added value or privacy concerns outweighs.

Outlook: Will AI replace editors in the future?

It is foreseeable that artificial intelligence will increasingly take over the extensive research and analysis before content creation. There are also more and more applications in content production where machines can replace human work. Bots already take over part of customer service by evaluating inquiries and providing relevant answers and content automatically. The fields of objective result reporting described above can also be covered by “robot journalists.”

But can artificial intelligence create every type of content automatically, without any human intervention? Let's consider the storytelling strategy. An advanced AI system will certainly be able to write much more than a matter-of-fact game report in the future. With the help of synonyms, metaphors, and other elements that make up natural human language to express emotions, a “machine” can also create content that tells a story, engages, and carries the reader with it. A step in this direction is the Watson system, which independently filled an entire issue of the British magazine "The Drum" with content and images.

It is certainly conceivable that machine learning could produce an expert article with an opinion, provided the system has enough information. However, when it comes to something that was personally experienced by a real person, it will probably be difficult to automate. Emotionally charged testimonials will therefore continue to be written by humans in the future.

In content creation, the following always applies: What decides the outcome is the data basis and the guidelines that people provide to the system, because a machine does not inherently have morals or the like. A specific political or moral orientation of a company must be learned by a system.

It is also questionable to what extent artificial intelligence can replace human creativity, emotions, and critical thinking. A complex algorithm can certainly imitate these typically human characteristics to a certain degree, but whether a machine can actually replace them equivalently in the foreseeable future is doubtful. And perhaps it will even be a quality feature of portals, blogs, or magazines in the future: "You can be sure that our content is written by real people!"