102 Turing’s Test
automated processes will require large government data sets becoming
more widely available, more sophisticated software, and a developing
digital economy that would allow organisations to make cost savings
through new, and initially expensive, technologies.
8
Since that paper’s
publication, most of these conditions are finally falling into place although it remains the case that in the case of investigative journalism,
software is still better placed in the service of human reporters, as we
shall see towards the end of this chapter. What Anderson (ironically)
referred to as “the future of news” in 2012 still remains, with regard
to sourcing many news stories, some way off,
9
but computer-generated
news production is beginning to take over in some fields. As Montal
and Reich point out in their survey of literature on algorithmic journalism, algorithms should no longer be thought of as precise recipes for
step-by-step instructions, but rather as computational actions that can
operate at high levels of uncertainty. It is this increased sophistication
of computational power that finally enables software to begin creating
automated journalism, which they define as “algorithmic processes that
convert data into narrative news texts with limited to no human intervention beyond the initial programming”.
10
Automated journalism is
finally with us, but just as the word “computer” originally referred to
humans carrying out iterative processes, so the algorithm of news is a
much older invention.
The comparison between Automated Insights and Josh Boak at the
start of this chapter is not intended to insult his work at all. Again and
again, Automated Insights – one of a number of companies such as Narrative Science using artificial intelligence to generate copy – makes it
clear that the AP style guide is a standard for their work, not simply
because of the business link between them but because (along with Reuters and the BBC) Associated Press is one of the globally recognised
standards for how to write journalistic prose. Concise sentences, direct
statements, judicious use of the “5 Ws”, and the news pyramid – anyone
who has ever taught how to write news will understand this approach:
very simply, there is a great deal of news writing that has been governed
by algorithms (or, as we have preferred to call them, style guides) for the
better part of a century. While much of the process of newsgathering (interviews, asking questions, dealing with truculent or awkward people,
and so on) cannot yet be done by machines, where highly structured,
machine-readable information exists, much of the writing of the news is
ripe for automation.
For the past 20–30 years, there has been a prevalent tendency among
some academics to view a particular model of journalism and news reporting as having been “invented” within America and the United Kingdom. Rather than tracing the fairly disparate roots of journalism to the
sixteenth and seventeenth centuries, authors such as Chelaby, Downing,
and Williams have shown that a particular format of writing became
automated processes will require large government data sets becoming
more widely available, more sophisticated software, and a developing
digital economy that would allow organisations to make cost savings
through new, and initially expensive, technologies.
8
Since that paper’s
publication, most of these conditions are finally falling into place although it remains the case that in the case of investigative journalism,
software is still better placed in the service of human reporters, as we
shall see towards the end of this chapter. What Anderson (ironically)
referred to as “the future of news” in 2012 still remains, with regard
to sourcing many news stories, some way off,
9
but computer-generated
news production is beginning to take over in some fields. As Montal
and Reich point out in their survey of literature on algorithmic journalism, algorithms should no longer be thought of as precise recipes for
step-by-step instructions, but rather as computational actions that can
operate at high levels of uncertainty. It is this increased sophistication
of computational power that finally enables software to begin creating
automated journalism, which they define as “algorithmic processes that
convert data into narrative news texts with limited to no human intervention beyond the initial programming”.
10
Automated journalism is
finally with us, but just as the word “computer” originally referred to
humans carrying out iterative processes, so the algorithm of news is a
much older invention.
The comparison between Automated Insights and Josh Boak at the
start of this chapter is not intended to insult his work at all. Again and
again, Automated Insights – one of a number of companies such as Narrative Science using artificial intelligence to generate copy – makes it
clear that the AP style guide is a standard for their work, not simply
because of the business link between them but because (along with Reuters and the BBC) Associated Press is one of the globally recognised
standards for how to write journalistic prose. Concise sentences, direct
statements, judicious use of the “5 Ws”, and the news pyramid – anyone
who has ever taught how to write news will understand this approach:
very simply, there is a great deal of news writing that has been governed
by algorithms (or, as we have preferred to call them, style guides) for the
better part of a century. While much of the process of newsgathering (interviews, asking questions, dealing with truculent or awkward people,
and so on) cannot yet be done by machines, where highly structured,
machine-readable information exists, much of the writing of the news is
ripe for automation.
For the past 20–30 years, there has been a prevalent tendency among
some academics to view a particular model of journalism and news reporting as having been “invented” within America and the United Kingdom. Rather than tracing the fairly disparate roots of journalism to the
sixteenth and seventeenth centuries, authors such as Chelaby, Downing,
and Williams have shown that a particular format of writing became