24 / 84
systems. The well-known historian,
Yu-val Harari, quotes the example of
Google’s AlphaZero, which took less
than 4 hours to learn how to play
chess, after which it was able to
beat the best human-trained chess
computer. Not by learning from
historical data, but by using machine
learning to teach itself to play the
game. This has gone so far now that
when players in chess tournaments
make a move that is strikingly
creative and original, the judges
may suspect that player is using a
chess computer to come up with the
moves. So, both routine jobs and
jobs in which new possibilities have
to be ex-plored can be carried out
by AI systems. But will they ever be
as crea-tive as we, humans? That
remains to be seen.
Whatever happens, the content of
many jobs will be changed by the
arrival of AI. We will work with these
AI systems and have to keep ad-
justing to new capabilities. Lifelong
learning will be very important, both
for low-skilled and highly qualified
jobs. New types of job will also be
created that we can’t predict at
the moment. Typically, jobs where
human contact is very important –
such as nursing – will still be done
by humans, even though robots
may be brought in to provide assis-
tance and support.
But the impact goes further than
just our job content and we will
also find other economic models
becoming involved in which we will
work less and in which job, income
and consumption will be separated
out from one another. So perhaps
you won’t have to work (fulltime) to
re-ceive a full income – and maybe
you won’t need a full income to pro-
vide for your basic needs. Just look
at the current digital economy of
apps and digital services. You can
use many services free of charge
and sometimes (with or without you
realizing it) in exchange for access
to your data. With this system, data
will become a new tradeable com-
modity and so we will be able to
ensure better that our data is our
own property and that we can use
it transparently when and where we
want.
How is imec working on this future?
Imec is working on neuromorphic
chips that are able to support
com-plex algorithms efficiently
and without consuming too much
energy. Being economical is
important when it comes to building
AI into sen-sors. Imec is also
involved in the ExaScience Life Lab
to build super-computers for major
medical problems (developing new
medication, understanding diseases
better, etc.).
The imec research group SMIT set up
the internal DANDA project to make
AI more transparent en to make
the algorithms understandable for
different parties involved. Methods
such as post-it data flow map-
ping and a description of the data
preparation process were developed
for imec’s AI developers. Also, the
DELICIOS project was launched in
2018. This four-year project will see
researchers examining which com-
plex tasks humans will be willing
to transfer to autonomous systems
and on what terms – as well as
how comfortable they feel about it.
Trusting these systems will be one
of the central points in the study.
A technology such as AI cannot
be developed without taking
account of the social and economic
implications from the outset. For
this reason, imec will be intensifying
this multidisciplinary research in
2019.
This article is part of a special
edition of imec magazine. To mark
imec’s 35th anniversary, we are
creating a vision of how technology
will shape our society in 2035.
Biography Pieter Ballon
Professor Pieter Ballon gained his
master’s degree in Modern History
from KU Leuven and his PhD in
Communication Sciences at VUB. He
has taught Communication Sciences
at VUB since 2009. Since 2016, he
has been director of SMIT, an imec
research group at VUB focusing on
‘Studies in Media, Innovation and
Technology’. Pieter Ballon was ap-
pointed the first Brussels Smart City
Ambassador and is also the Inter-
national Secretary for the European
Network of Living Labs. His publica-
tions include the book “Smart Cities:
how technology keeps our cities
livable and makes them smarter”.
Illustrations for article pieter ballon:
in attach a picture that is ours. The
other are from istock and one from
MIT (request picture usage from
norman project with Pinar Yanardag
Delul
pinary@mit.edu)24 l New-Tech Magazine Europe