AI & Content Creation
AI is already creating content online, often used by Associated Press to craft simple, statistics-based copy about things like sport and finance, where numbers and results are key. In fact, Aaron Souppouris of Engadget recruited his very own bot to write an announcement-style blog looking at the release of the Samsung Galaxy Note 7 smartphone.
He did this by training his bot to write based on hundreds Engadget’s previous articles. Editing variables and using a synonym tool allowed the bot to come up with a range of different versions of the same paragraph, choosing its own variations to say the same thing. The article itself, while quite dry, still made sense and gave the reader an accurate comparison of the new Samsung phone with rival products.
Of course, there is an element of art vs science about teaching machines to write copy. There is a lack of emotion and empathy in computer-generated words, as well as a loss of the flexibility that is associated with content writing. For example, any pop culture references, which so often help us identify with a topic or ideology, will be noticeably absent from AI-generated content.
Although neural networks are developing all the time, they have yet to simulate the purity of language that comes from a human mind. Whilst number-based pieces containing easily defined variables can be created using AI, the heart of a piece still needs to come from, well, the heart.