120 Turing’s Test
in the first chapter of this book, general artificial intelligence remains
a grail that is often the stuff of legend rather than reality. As such, the
issue of algorithmic journalism, of robots versus humans, is again much
more accurately considered a topic of automation rather than real AI. In
the field of journalism and reporting, there are many types of cognitive
work that are repetitive and iterative, that need to be presented in a consistent way according to clear criteria and source of data, the processes
of which can be routinely and effectively described by algorithms. In
the fields of sports and financial reporting, particularly for minor sports
or companies where there simply are not enough – and have never been
enough – human journalists to cover those activities, and where information is clearly structured and readily available, people will not be able
to compete against software in domains where people are looking for information, for data, provided in a clear narrative form. And yet, as with
the example of Eric Eyre but also the countless journalists reporting on
Donald Trump, fundamentals such as determining what the data actually are and then collating them from multiple, often unreliable, sources
into something that is coherent appear to be a task that, for the foreseeable future, is beyond the capacity of machines.
Not that Eyre’s strengths as a reporter have been an unmixed blessing
to the Charleston Gazette-Mail: in 2015, two years before receiving his
Pulitzer, he along with other journalists had to reapply for his job when
the Gazette and the Daily Mail merged, while in 2018 the company that
owned the newspaper filed for bankruptcy, eventually selling the title to
HD Media in March. As with so many other areas, the Gazette-Mail
was struggling to compete in an environment where both sales and advertising revenues were declining rapidly. The role of technology, then,
is very often not a question of capabilities of human versus algorithmic
journalists, but rather one of costs and finances. Technical innovation is
frequently driven by such constraints, and the impetus for algorithmic
journalism is frequently to drive down costs by replacing peoples’ salaries with software subscriptions. In the analytics-driven world of contemporary journalism, traffic is all and some of the wiser producers such
as AP have begun to invest in the “long tail” of audiences, a mass production of hundreds of thousands of stories that will only capture a small
number of readers each but which aggregate into significant numbers
overall. Within such a difficult environment, however, there are significant challenges for extending automated journalism. Associated Press
and other organisations will certainly begin to use algorithmically generated content for health, crime, some social reporting, and other elements
where clear data are publicly available, but any form of investigative
journalism appears beyond the abilities of even the best software for the
foreseeable future. Yet, to repeat, this is not some simplistic rendition of
human versus machine: in those areas where algorithms are not capable
of generating content, what we more often see is augmented journalism.
in the first chapter of this book, general artificial intelligence remains
a grail that is often the stuff of legend rather than reality. As such, the
issue of algorithmic journalism, of robots versus humans, is again much
more accurately considered a topic of automation rather than real AI. In
the field of journalism and reporting, there are many types of cognitive
work that are repetitive and iterative, that need to be presented in a consistent way according to clear criteria and source of data, the processes
of which can be routinely and effectively described by algorithms. In
the fields of sports and financial reporting, particularly for minor sports
or companies where there simply are not enough – and have never been
enough – human journalists to cover those activities, and where information is clearly structured and readily available, people will not be able
to compete against software in domains where people are looking for information, for data, provided in a clear narrative form. And yet, as with
the example of Eric Eyre but also the countless journalists reporting on
Donald Trump, fundamentals such as determining what the data actually are and then collating them from multiple, often unreliable, sources
into something that is coherent appear to be a task that, for the foreseeable future, is beyond the capacity of machines.
Not that Eyre’s strengths as a reporter have been an unmixed blessing
to the Charleston Gazette-Mail: in 2015, two years before receiving his
Pulitzer, he along with other journalists had to reapply for his job when
the Gazette and the Daily Mail merged, while in 2018 the company that
owned the newspaper filed for bankruptcy, eventually selling the title to
HD Media in March. As with so many other areas, the Gazette-Mail
was struggling to compete in an environment where both sales and advertising revenues were declining rapidly. The role of technology, then,
is very often not a question of capabilities of human versus algorithmic
journalists, but rather one of costs and finances. Technical innovation is
frequently driven by such constraints, and the impetus for algorithmic
journalism is frequently to drive down costs by replacing peoples’ salaries with software subscriptions. In the analytics-driven world of contemporary journalism, traffic is all and some of the wiser producers such
as AP have begun to invest in the “long tail” of audiences, a mass production of hundreds of thousands of stories that will only capture a small
number of readers each but which aggregate into significant numbers
overall. Within such a difficult environment, however, there are significant challenges for extending automated journalism. Associated Press
and other organisations will certainly begin to use algorithmically generated content for health, crime, some social reporting, and other elements
where clear data are publicly available, but any form of investigative
journalism appears beyond the abilities of even the best software for the
foreseeable future. Yet, to repeat, this is not some simplistic rendition of
human versus machine: in those areas where algorithms are not capable
of generating content, what we more often see is augmented journalism.