4 / 20
A new steam engine
The internet is probably as important to the development of
artificial intelligence as the coal-fired steam engine was to the
development of the coal industry.
The American author and vlogger John Green explains in an
interesting way how the industrial revolution gathered speed.
The invention of the steam engine made coal itself important to
the coal industry, as coal-fired steam engines made it cheaper
and easier to extract more coal. More coal in turn laid the
foundation for more coal-fired steam engines. Railways were
important to the steel industry in the same way; railways made
it much easier to transport heavy steel, including rails, and this
made it easier to extend the railways and to extract even more
steel. John Green thus provides a good illustration of what
exponential growth looks like in practice.
Researchers and analysts believe artificial intelligence is central
to the fourth industrial revolution. Artificial intelligence is not a
new concept, having arisen as long ago as the 50s. For a long
time, the concept of artificial intelligence was primarily associ-
ated with science fiction. In recent years, however, things have
moved on, as exemplified by driverless cars, voice-operated
applications, and voice and image recognition technologies.
Try, for example, something as simple as taking five group
photos with your phone and uploading them to Google
Photos. Google will ’look’ at your photos, and will probably,
without you even having asked it to, create a sixth photo that
combines the best smiles from the five existing photos – and
no-one will have their eyes shut in the sixth photo either. In
Google Photos you can also search for mountains, cars, boats,
seas, palm trees, dogs etc. to help you find photos of specific
things you have captured with your camera.
The internet provides access to a nearly limitless quantity of
unstructured information. This, coupled with greater processing
capacity and the development of what are known as neural
networks, is making it possible for machines to sort information
through different layers in a network.
For example, photos can be sorted in a first layer according
to whether they are of a living thing or not. They can then be
sorted bywhether they are of an animal or a human (if a living
thing), or bywhether they are of scenery or objects. Search
results and other programmed responses enable such neural
Foreword
EVRY SURVEY
4