DATA-DRIVEN INNOVATION and the implications on jobs and skills [email protected] http://oe.cd/bigdata Working Group on Legal Questions related to the Development of Robotics and Artificial Intelligence 26 May 2015
DATA-DRIVEN INNOVATION and the implications on jobs and skills
[email protected] http://oe.cd/bigdata
Working Group on Legal Questions related to the
Development of Robotics and Artificial Intelligence
26 May 2015
DDI refers to the use of data and analytics
to improve or foster new products,
processes, organisational methods and
markets
2
What is Data-Driven Innovation (DDI)?
Big data feeding ML algorithms to
enable autonomous decision making
3
Value added
growth
and well-being
Knowledge base
Decision
making
Datafication and
data collection
Big data
Data value cycle
Data analytics
(machine learning, ML)
DDI enables next generation
autonomous machines and systems
Manufacturing
Agriculture
Finance 4
Logistic
Health
Transportation
5
Example: algorithmic trading in finance
Algorithmic trading as share of total trading
Note: 2013-14 based on estimates.
Source: OECD based on The Economist (2012) and Aite Group
• Internet of Things (IoT) – embedding physical
objects in data flows (and intelligence)
– Driverless cars enabled by information flows from road
infrastructure, other cars, and web services
• IoT empowering, but also embedding humans in
data flows
– “The IoT requires thinking about how humans and things
cooperate differently when things get smarter.” (Tim
O’Reilly)
– Leading to the emergence of an intelligent
“superorganism”? 6
IoT will be the next game changer
7
Towards the next production revolution?
1. Who will be the losers and winners in the
“race against the machines”?
2. Do we have the capacity to “dance with
the machines”?
3. What are the challenges faced by the
human “dancer”?
8
What are the employment implications?
WHO WILL BE THE LOSERS AND WINNERS?
10
We have to learn from history!
Jacquard loom punch cards
Handmade damasks
Mechanical tabulator
Frey and Osborne (2013)
• Creative intelligence
• Social intelligence
• Complex perception and manipulation
Levy and Murnane (2013)
• Solving unstructured problems
• Working with new information
• Non-routine manual tasks
Elliott (2014)
• Language reasoning
• Vision movement
11
Capacities needed to successfully
“race against the machines”
12
Solving unstructured problems and
working with data will be key!
Index of Changing Work Tasks in the U.S. Economy
Source: Levy and Murnane, 2013
13
Implications on inequalities …
Trends in wages for full-time, full-year male workers in the United States,
1963-2008
Source: Brynjolfsson and McAfee, 2014 based on Acemoglu and Autor (2011)
14
… with the share of income
going to labour declining steadily.
0.55
0.60
0.65
0.70
0.75
0.80
1980 1990 2000 2010
Australia Germany Japan United States
Source: OECD Unit Labour Costs – Annual Indicators
15
What used to be attributed to labour is
now knowledge-based capital.
• IP
• Software
(e.g. ERP,
algorithms)
• Data (“Big”)
• Creativity
• Expert decision making
• Organisational know-how
• Marketing, sales, customer
relations
Capital Labour
Ownership of autonomous machines and
systems will be defined by IP rights
DO WE HAVE THE CAPACITY TO
“DANCE WITH THE MACHINE”?
Poor ICT adoption in many businesses!
The diffusion of selected ICT tools and activities in enterprises, 2013
Percentage of enterprises with ten or more persons employed
Source: OECD, ICT Database; Eurostat, Information Society Statistics and national sources, July 2014.
17
FINFIN
NZL
ISL
SWEPRT FIN
NZL
KOR
GRC
TUR
TUR CZEGBR HUN
POL
ITA
GBR0
20
40
60
80
100
Broadband Website E-purchases Social network ERP Supply chainmngt. (ADE)
Cloud computing E-sales RFID
%
Highest Lowest 1st and 3rd quartiles Median Average
Knowledge-based capital related workers, 2012
(as a percentage of total employed persons)
18
Organisational change needs to be
encouraged
Source: OECD Science, Technology and Industry Scoreboard 2013.
http://dx.doi.org/10.1787/888932890618
0
5
10
15
20
25
30
TUR SVK ITA PRT HUN ESP GRC DNK POL CZE LUX AUT IRL FIN SVN EST BEL NLD SWE DEU FRA NOR ISL GBR USA
Organisational Capital Computerised Information Design Research & Development Overlapping assets
19
Poor proficiency in problem solving in
technology-rich environments
100
80
60
40
20
0
20
40
60
80
100
Level 1 or below Level 2 Level 3
No ICT skills or basic skills to fullfilsimple tasks
More advanced ICT and cognitive skills to evaluate
problems and solutions
Source: OECD Science, Technology and Industry Outlook 2014, based on OECD’s Programme for the
International Assessment of Adult Competencies (PIAAC), http://dx.doi.org/10.1787/888933151932.
As a percentage of 16-65 year-olds (2012)
20
Get the basic skills right!
05
10
15
20
25
30
FIN
NZ
L
JP
N
AU
S
DE
U
NL
D
CA
N
KO
R
GB
R
CH
E
ES
T
BE
L
SV
N
US
A
IRL
OE
CD
CZ
E
FR
A
SW
E
AU
T
PO
L
ISL
DN
K
LU
X
NO
R
SV
K
ITA
HU
N
RU
S
PR
T
ES
P
ISR
GR
C
TU
R
CH
L
BR
A
ME
X
IDN
%
Scie
nce
Re
ad
ing
Ma
the
ma
tics
Science, reading and mathematics proficiency at age 15, 2009
OECD Science, Technology and Industry Scoreboard 2013, http://dx.doi.org/10.1787/888932890675
based on PISA 2009 Results: What Students Know and Can Do: Student Performance in Reading,
Mathematics and Science, Vol. 1, OECD Publishing.
• “… it’s technology married with liberal arts, married with the humanities, that yields us the results that make our heart sing” (Steve Jobs)
• STEM need to be complemented with a broader interdisciplinary understanding of multiple complex subjects (e.g. legal & ethics)
• People need to rediscover their bodies as highly developed sensomotoric skills will also become a key competitive advantage
21
Science, technology, engineering and
mathematics (STEM) are not enough !
WHAT ARE THE CHALLENGES DECISION MAKERS
WILL FACE ?
Automated decision-making is not
perfect!
23 Source: Nature.com
• Need for enhancing
the transparency of
automated decisions
in some areas.
24
How to improve the transparency of
algorithms?
• However, transparency efforts need to respect the IPRs (incl. trade secrets) of businesses.
• Where a machine contradicts the opinion of
the human decision maker, will [s]he be willing
and able to take over the responsibility when
overriding the machine’s suggested decision?
• Risk of a “dictatorship of data”, where less
educated/concerned decision makers
automatically follow the decisions of machines
25
Need for clearer accountability and
responsibility assignments
26
Thank you for your attention!
• Source: Chapter 6 of “Data-driven Innovation: Big Data for Growth and Well-being”
• To be presented at the OECD Forum in 2-3 June 2015
• To be released in September 2015
• http://oe.cd/bigdata
Contact: Christian.Reimsbach-Kounatze [aatt] oecd.org