The future of news is already written

It’s fair to say that journalism is in crisis and much of the blame can be ascribed to the rise of the Web. Since the 90s, advertising revenues have followed a dramatic descent and the traditional newspaper business model dependent on that revenue is under existential threat. Nevertheless, technology which has brought about an established industry’s near destruction may also offer a way forward, according to Jim Kennedy, Vice President of strategy and development at the Associated Press, the largest news-gathering operations in the world.

In 2008, Kennedy was responsible for the creation of a department devoted to launching digital products, the most significant and latest of which is harnessing artificial intelligence to write stories. Following a definitive proof of concept trial run in 2013 which took the data stream from stats provided on the weekly performance of NFL offensive players to create short stories, production shifted in 2014 to other news platforms; earnings stories on publicly-held companies.

From an average of approximately one hundred 150 word articles per month on the subject based on the output of AP’s 65 business reporters, production increased by a factor of 12  to end in a total of 3700 earnings stories per quarter by the end of 2015.

The important point here is that a machine has managed to perform a task which has otherwise been reserved for humans alone with far greater efficiency and so well as to be indistinguishable from that written by humans. While this is still a distant parallel to the test developed by Alan Turing in 1950 to gauge a machine's ability to exhibit intelligent behaviour equivalent to, or indistinguishable from that of  a human, it is nevertheless, a watershed of its kind.

Kennedy is keen to point out that the aim of using AI-generated content is not an attempt to replace journalists but rather to supplement their work and provide greater coverage to local readerships on stories that would otherwise not have been written.

Not only this but AI will also help journalists to gain insight from the vast amount of data available from organisations, people and issues as well as enable readers to drill down information in articles based on tags. This last function Kennedy believes offers consumers of AP’s products an extra level of convenience highly suitable to the on-demand consumer world.

It is also his contention that the abundance of stories arising from AI production will lead to further abundance; journalists will be freed from stand reporting stories to focus on more complex articles that are otherwise time consuming and difficult to produce.

And news is nevertheless a product for consumption, even as the platforms where it is consumed have changed. A recent article in the online publication Wired explained that news is driving engagement in the social media space more than ever, with Facebook hosting more than one billion ‘media interactions’ per day.

In a highly volatile media space characterized by an almost insatiable need for fresh content, the introduction of AI is a potential game changer. Publishers have finite resources and limited time but machines, once they have mastered the means can produce content with tireless efficiency and speed. Whether this adds to the value of human-generated content, or whether the two eventually become indistinguishable is a matter for science fiction writers to ponder – for now.

If you liked this article, please feel free to review the latest edition of Perspectives by TCS on the subject of AI and other compelling themes.