Jason Dooris is the CEO and founder of growing Atomic 212, Australia's fastest growing media and marketing agency on the BRW 2014 list. Over the past 20 years, Jason has held a variety of senior local and global industry positions including CEO MediaCom UK, deputy CEO MediaCom Europe, GM Saatchi & Saatchi NZ, GM Ogilvy & Mather Australia, GM Dentsu Aegis Australia and consulting practice director, Deloitte Asia. His vertical experience covers most categories with a particular focus on retail, automotive and FMCG.
Computers and artificial intelligence have come along at an
exponential rate over the past few decades, from being regarded as
oversized adding machines to the point where they have played integral
roles in some legitimately creative endeavours.
Think Microsoft’s project ‘The Next Rembrandt’,
where computers were used to analyse all 346 of the Dutch master’s
works, before an entirely new ‘painting’ was 3D printed (so it would
have texture and brushstrokes). The result was a piece the untrained eye
would still say was absolutely an original Rembrandt.
Or Flow Machines, from Sony CSL research laboratory, which
analysed 13,000 leadsheets of music to create a little ditty by the name
of ‘Daddy’s Car’. It’s a pretty decent copy of something The Beatles
might have whipped up around their Sgt. Pepper's era.
They’re the kind of advancements in AI that, on face value, might
have the world’s creatives feeling slightly nervous. Sure, robots might
be on track to replace repetitive, manual labour jobs, but original
thought is surely a distinctly human trait – are we really this close to
SkyNet becoming sentient?
You can breathe a sigh of relief, because while machines are
making impressively close copies of both Rembrandt paintings and Beatles
songs, there’s still a huge human element involved.
In the case of The Next Rembrandt, while computers technically
did the painting – more accurately, a 3D printer did – a team of
computer engineers were behind the entire project, and spent 18 months
putting the whole thing together. As for Flow Machines, human composer,
Benoît Carré, decided on the style, created the leadsheet for the song,
wrote the lyrics, and did the production and mixing.
But can we afford to be complacent about our superiority over
machines? Is it on the cards that our place in the ideation process
could be superseded by AI?
A decades-old debate
This is more than just a question we’ve been pondering since we
could first ask Siri what to wear when we go outside today. People have
been grappling with the idea of computational creativity since before
the advent of the personal computer.
In 1982, Marvin Minsky, co-founder of the Massachusetts Institute of Technology's AI laboratory, wrote a piece entitled ‘Why people think computers can’t’, which largely discussed this issue.
One of the first ideas Minksy brings up is that we should dismiss
the temptation to think of creativity in terms of Einstein or
Beethoven. He says: “I don't believe that there's much difference
between ordinary thought and highly creative thought.”
Regarded as one of the great minds in cognitive science, AI and
philosophy, Minksy’s paper delves into such abstract ideas as the nature
of self-awareness and the mind. But he keeps things basic and easy for
those of us who aren’t Turing Award recipients.
Minsky also argues that because we don’t really understand how
‘creative’ thoughts come about, we tend to “forget how little we know
about the marvels of ordinary thinking”.
So perhaps it’s a good idea to focus on the notion that a
computer can ‘think’ rather than just be ‘creative’ – achieving the
first would surely lead to computers attaining the second.
Is your job in trouble?
A 2013 study by University of Oxford researchers Carl Benedikt Frey and Michael Osborne made a hugely publicised estimation that 47 per cent of jobs in the US could be done by a robot or AI by the year 2033.
It understandably caused some serious waves; almost half of one
of the largest workforces in the world being out of a job is an alarming
stat. But what was less publicised was the final line, which stated
“For workers to win the race, however, they will have to acquire
creative and social skills”.
The actual findings in the paper were that 47 per cent of the
workforce were at ‘high risk’ of losing their job to an automated
system, while those at the lowest risk were people working in
“generalist occupations requiring knowledge of human heuristics, and
specialist occupations involving the development of novel ideas and
Why does a computer need to be creative?
It’s a question that often lies at the heart of whether computers can create: Why would they bother?
As The Next Web’s former
editor-in-chief, Matthew Hussey, put it, “A computer doesn’t deal with
grief, love, pain, sadness, sorrow or any of the infinite combinations
of emotion that make up the rich tapestry of our lives. We have
necessity to be creative, computers, simply don’t.” And at first glance,
that does seem like the definitive argument that computational
creativity is merely a fiction.
But Minsky dealt with this pretty definitively way back in 1982.
To those who claim computers aren’t capable of possessing what we would
consider a ‘mind’ because that requires feelings or thoughts, Minsky
suggested you shift your thinking to say: “Computer can't do [XX],
because all they can do is execute incredibly intricate processes,
perhaps millions at a time.”
While we are perhaps in the dark as to our brains’ full capabilities – although that whole ‘only works at 10 per cent of its capacity’ is bull
– watching even a smartphone do its thing is a reminder that computers
have insane capabilities these days, perhaps beyond what anyone could
have conceived in 1982.
So to dismiss out of hand the idea that computers could ever have ideas seems a foolish statement.
But if Carl Benedikt Frey and Michael Osborne are on the money
with their study, if your job involves coming up with ideas and solving
human problems, you’ve got at least two decades of employment security
when it comes to robots.