116 Turing’s Test
very much limited to open government data sets. Nonetheless, as she
also observed, “for those communities that have lost these local reports
on crime or unemployment when their local paper shut down, services
such as RADAR can help fill the gap.”
63
Algorithmic journalism as yet
lacks the flexibility to work across a vast range of topics that cannot
automatically be structured to clearly defined templates, but in many
places as news is gutted by a rapidly changing and often hostile commercial environment, the choice is increasingly not between robot journalists and human journalists, but between robot journalists and no
journalists at all.
Augmented Journalism
As has previously been noted, in 2013, automated software produced
300 million pieces of content, more than all legacy media companies
combined according to Ulanoff. More recent specific figures outside of
clients such as Associated Press, who are very public about both which
articles are produced by software and how many are published at any
time, are more difficult to ascertain; many other organisations that use
Wordsmith are less forthcoming about the fact that their content is produced by an algorithm. Assuming that clients include realtors and sports
sites as well as Amazon, Ai’s more recent claim that it has “written billions of articles” may be true. With such a vast quantity of content being
produced algorithmically and set to move into the field of local reporting, it would seem that algorithmic journalism is not the future but has
already arrived. Yet before rushing to consider a future dominated by
robot reporters, a few other factors should be borne in mind before we
consign people to the dustbin of journalism. One of those factors is another piece of electronic communication in which, in terms of quantity
alone, machines appear to reign supreme: email.
According to the research group Radicati, global email traffic at the
beginning of 2018 was estimated at 281.1 billion emails per day being
sent between over 3.8 billion accounts, and predicted to rise to more
than 333 billion messages per day by 2022.
64
There are no reliable statistics for the quantity of daily messages which are spam but, according
to Statista, in March 2018, more than 48 per cent of emails were spam.
65
As a great deal of spam is automatically generated by spambots – a
very crude version of the algorithmic content generation employed by
Wordsmith – it would not be unreasonable to presume that software is
producing billions of messages daily, the vast majority of which is never
read by its intended audience because it is filtered out. Automatically
generated content is a significant portion of daily messages transmitted
across the globe, but that does not mean that it is automatically read.
The work done by Associated Press and Automated Insights is very, very
far from the crudely produced content seeking to promote adult goods
very much limited to open government data sets. Nonetheless, as she
also observed, “for those communities that have lost these local reports
on crime or unemployment when their local paper shut down, services
such as RADAR can help fill the gap.”
63
Algorithmic journalism as yet
lacks the flexibility to work across a vast range of topics that cannot
automatically be structured to clearly defined templates, but in many
places as news is gutted by a rapidly changing and often hostile commercial environment, the choice is increasingly not between robot journalists and human journalists, but between robot journalists and no
journalists at all.
Augmented Journalism
As has previously been noted, in 2013, automated software produced
300 million pieces of content, more than all legacy media companies
combined according to Ulanoff. More recent specific figures outside of
clients such as Associated Press, who are very public about both which
articles are produced by software and how many are published at any
time, are more difficult to ascertain; many other organisations that use
Wordsmith are less forthcoming about the fact that their content is produced by an algorithm. Assuming that clients include realtors and sports
sites as well as Amazon, Ai’s more recent claim that it has “written billions of articles” may be true. With such a vast quantity of content being
produced algorithmically and set to move into the field of local reporting, it would seem that algorithmic journalism is not the future but has
already arrived. Yet before rushing to consider a future dominated by
robot reporters, a few other factors should be borne in mind before we
consign people to the dustbin of journalism. One of those factors is another piece of electronic communication in which, in terms of quantity
alone, machines appear to reign supreme: email.
According to the research group Radicati, global email traffic at the
beginning of 2018 was estimated at 281.1 billion emails per day being
sent between over 3.8 billion accounts, and predicted to rise to more
than 333 billion messages per day by 2022.
64
There are no reliable statistics for the quantity of daily messages which are spam but, according
to Statista, in March 2018, more than 48 per cent of emails were spam.
65
As a great deal of spam is automatically generated by spambots – a
very crude version of the algorithmic content generation employed by
Wordsmith – it would not be unreasonable to presume that software is
producing billions of messages daily, the vast majority of which is never
read by its intended audience because it is filtered out. Automatically
generated content is a significant portion of daily messages transmitted
across the globe, but that does not mean that it is automatically read.
The work done by Associated Press and Automated Insights is very, very
far from the crudely produced content seeking to promote adult goods