The academic world recently learned that bots can write automated gibberish and — with a little help from their fleshy minions — can have it published in mainstream peer-reviewed scientific publications. But are we prepared for what follows from the moment when bots can reliably produce writing that makes real sense and which is useful and timely enough for use in major newspapers? It’s happening already. The finances of newspapers are such that a wave of robo-journalism seems inevitable, once we have a few more advances in semantics and automated basic fact-checking. Given the current dismal state of newspaper science reporting such new-fangled robo-news may even be slightly better than what we have now.
It follows that journal editors and publishers may soon need to add a new clause to their author guidelines, such as: “articles must be fully written by humans”. Not for fear of gibberish faux-papers, but rather because bots will be able to add sensible summaries and otherwise usefully aid in the writing of a research paper. Or we may need to develop an agreed form of simple presentation to flag up:
[bot]this section of the text was written by bots[/bot] and to embed links to the bot’s sources.
Incidentally, I’ve also often thought that the humourous LOLcat language would form a pleasing basis for identifying messages-sent-to-humans by objects embedded in The Internet of Things, clearly marking their simpler forms of communications to us as being: ‘not kreated by th humanz’. We already have the LOLcat translation systems available.