Automated journalism: A posthuman future for digital news?

Whether or not a human author wrote these words is not a question that most readers will consciously ask. Writing has been an unambiguously human activity, one whose durability has persisted through the transition from the intimacy of the written word to the mass reproducibility of print to the explosion of digital media. Even as human-computer interaction has become normalized in daily life (Reeves and Nass, 1996), the creative act of composition has appeared cordoned off as something innately human. The very root of ‘author’ in the Latin auctor-creator-and its connection to auctoritas-a kind of authority (Höpfl, 1999)-bind humanness to the written word. Yet such close ties can no longer be assumed. Advances in artificial intelligence and natural language generation have moved nonhuman writing from theory to practice, with journalism at the forefront of industries being affected. Already, human-authored news stories are being joined by a growing number of stories created entirely by computer programs. Just one company alone, Automated Insights, announced on its web site it has surpassed the milestone of a billion stories authored annually. While a billion seems incredible, the company’s current software has the capacity to produce 2,000 unique stories every second-or 63 billion stories annually. These numbers point not only to the qualitative shift from human to computer but also to a quantitative transformation that is difficult to comprehend when we still think of human journalists laboring over their computer screens. Accepting that machines work faster than humans is deeply embedded in understandings of technological innovation and contemporary labor in industrial applications; now these forces are descending on creative practices as well.