118 Turing’s Test
Stephen Hawking once famously observed in A Brief History of Time
that “each equation I included in the book would halve the sales”, and
on the surface of it Eyre’s news story would seem daunting to most readers. While the equation (780 million pills equal 1,728 deaths) is implicit
rather than explicit, the use of statistics and pharmaceutical terminology
offers no concessions. Yet far from alienating readers, these stories were
part of a transformative narrative within Americans that influenced the
government and began a far-reaching debate into an epidemic responsible for a rising number of deaths. As with stories produced by Ai, Eyre
was reliant on data for his articles, but at that point, the similarities
between the two methods of journalism cease.
Joseph Burns in a piece for the Association of Health Care Journalists
noted that the power of Eyre’s work came, at least in part, from “hard
data” and “big numbers”, providing strong evidence as to the scale of
the flood of opiates into depressed West Virginia counties.
69
Such information was important: while facts themselves do not automatically
constitute truth, as we have noted in the previous chapter, any interpretation of events that ignores or refutes such facts can never aspire to
be a truthful interpretation. Much of the authoritative nature of Eyre’s
journalism came from his careful use of a range of hard data. Yet, as
he made clear in an interview with Arka Gupta, simply collecting that
information reveals a problem for any form of investigative journalism
that many reporters will recognise immediately: the unwillingness of
those who hold data to make it public. The investigation actually began with a tip off in 2013 that the state Attorney General was involved
in a lawsuit against drug wholesalers, the details of which (along with
another case in 2016) were sealed at the request of both the companies
and the Attorney General. Intervening in the latter case, Eyre and the
Charleston Gazette-Mail argued that the information should be made
publicly available, which, in turn, led him to file a number of freedom
of information access requests to the office of the Attorney General in
order to release information from the Drug Enforcement Agency.
70
To
make matters more complicated, as Eyre told Linda Wertheimer, the
Gazette-Mail was a small newspaper with a circulation of some 40,000
copies, meaning that he had to put the story aside from time to time to
work on the rest of his day job.
71
While robot journalists can churn out
millions of articles that will effortlessly convert statistics into quarterly
reports, there are none in operation yet that can go to a court of law to
ensure that data are made publicly available in the first place.
Algorithmic journalism can only deal with information that is made
available (and then only if it has been structured in a machine-readable
format, rather than countless letters, notes, and recordings), which remains a fundamental problem for software when it comes to finding stories. In terms of the raw numbers game, robots win every time – within
a very limited field of operations. Yet, what is also outstanding about
Stephen Hawking once famously observed in A Brief History of Time
that “each equation I included in the book would halve the sales”, and
on the surface of it Eyre’s news story would seem daunting to most readers. While the equation (780 million pills equal 1,728 deaths) is implicit
rather than explicit, the use of statistics and pharmaceutical terminology
offers no concessions. Yet far from alienating readers, these stories were
part of a transformative narrative within Americans that influenced the
government and began a far-reaching debate into an epidemic responsible for a rising number of deaths. As with stories produced by Ai, Eyre
was reliant on data for his articles, but at that point, the similarities
between the two methods of journalism cease.
Joseph Burns in a piece for the Association of Health Care Journalists
noted that the power of Eyre’s work came, at least in part, from “hard
data” and “big numbers”, providing strong evidence as to the scale of
the flood of opiates into depressed West Virginia counties.
69
Such information was important: while facts themselves do not automatically
constitute truth, as we have noted in the previous chapter, any interpretation of events that ignores or refutes such facts can never aspire to
be a truthful interpretation. Much of the authoritative nature of Eyre’s
journalism came from his careful use of a range of hard data. Yet, as
he made clear in an interview with Arka Gupta, simply collecting that
information reveals a problem for any form of investigative journalism
that many reporters will recognise immediately: the unwillingness of
those who hold data to make it public. The investigation actually began with a tip off in 2013 that the state Attorney General was involved
in a lawsuit against drug wholesalers, the details of which (along with
another case in 2016) were sealed at the request of both the companies
and the Attorney General. Intervening in the latter case, Eyre and the
Charleston Gazette-Mail argued that the information should be made
publicly available, which, in turn, led him to file a number of freedom
of information access requests to the office of the Attorney General in
order to release information from the Drug Enforcement Agency.
70
To
make matters more complicated, as Eyre told Linda Wertheimer, the
Gazette-Mail was a small newspaper with a circulation of some 40,000
copies, meaning that he had to put the story aside from time to time to
work on the rest of his day job.
71
While robot journalists can churn out
millions of articles that will effortlessly convert statistics into quarterly
reports, there are none in operation yet that can go to a court of law to
ensure that data are made publicly available in the first place.
Algorithmic journalism can only deal with information that is made
available (and then only if it has been structured in a machine-readable
format, rather than countless letters, notes, and recordings), which remains a fundamental problem for software when it comes to finding stories. In terms of the raw numbers game, robots win every time – within
a very limited field of operations. Yet, what is also outstanding about