R&I enabling artificial intelligence

• Author: Ana Correia and Irina Reyes
• Pages: 451-509

451

CHAPTER 7

R&I ENABLING

ARTIFICIAL

INTELLIGENCE

KEY FIGURES

60 %

of all AI science

is in f‌ields other

than computer

science

19 %

EU28’s share in

world AI f‌irms

22 %

EU share

in global

AI publications

>€20 bn

per year of EU private and

public investments over

the next decade

8 %

EU28’s share of global

AI private investments

452

What can we learn?

 AI is a potential game changer for pro-

ductivity and sustainability, providing the

right complementary skills, infrastructure and

management culture are in place.

 R&I solutions are needed to mitigate the

environmental footprint of AI.

 AI is a vital tool in the f‌ight against the

new coronavirus. At the same time, the use

of AI tracking and surveillance tools in the

context of this pandemic has shown the need

for global ethical governance of AI.

 Data explosion, stronger computational

power, more sophisticated algorithms

and open source soware have enabled

breakthroughs in AI R&I.

 ‘AI dynamics’: exploring the boundaries of

scientif‌ic f‌ields beyond computer science, with

intersectoral and intensif‌ied cross-country

collaboration, EU included.

 The EU ranks among global leaders in

AI science but trails in AI innovation, al-

though it is in line with its share in global R&D.

 A gender diversity gap in AI research

persists but is less pronounced in Europe

than in other regions worldwide.

 Private investments and acquisitions

of AI startups are on the rise. EU

investments remain insuf‌f‌icient. The

United States leads, followed by China.

 AI talent is relatively scarce worldwide

and appears more predominant in the United

States than in the EU.

 AI is increasingly blending with digital

technologies, such as blockchain, and with

the physical world in f‌ields like advanced

manufacturing and materials science.

What does it mean for policy?

 AI can play a big role in the economic,

social and ecological transition Europe

is undergoing.

 The EU should capitalise on its scientif‌ic

and industrial strengths to lead in AI

development and to foster technologies

that both benef‌it and augment its potential.

 The EU and Member States need to join

forces to raise the level of public and pri-

vate investments in AI, deepen the Digital

Single Market, move towards AI technology

sovereignty, and dif‌fuse AI practices across

the Union.

 The EU needs to promote AI talent

production and retention in the EU (while

attracting foreign talent), investments

and capacity-building in related digital

technologies, such as high-performance

computing, European cloud and micro-

electronics, and research and digital

infrastructure, notably 5G.

 The EU’s guiding principles of trustworthy,

human-centric, and ethical AI are

a strength and not an obstacle to the

EU AI innovation ecosystem. These

will also improve the ‘trust in tech’ and

safeguard privacy.

453

CHAPTER 7

1. Artif‌icial intelligence: a potential game changer

for productivity and sustainability

1 Based on European Commission (2019), Report by the High-Level Expert Group on AI set up by the European Commission.

Artif‌icial Intelligence (AI) as a f‌ield of

study is already 70 years old. In 1950, Alan

Turing put forward the so-called ‘Turing test’ as

a way of determining if a computer is capable

of thinking like a human. John McCarthy,

a computer scientist, then coined the term

‘artif‌icial intelligence’ during a conference in

1955. Between 1955 and 1997 – when IBM’s

Deep Blue defeated Gary Kasparov, world chess

champion – there were periods of progress

in the f‌ield, oen restricted to highly specif‌ic

applications and notably in natural language

processing and neural networks. However,

there were also periods known as ‘AI winters’,

brought about by overly big expectations, a lack

of practical applications of AI and, ultimately,

reductions in AI research funding. In 2006,

developments in deep learning generated

further enthusiasm around AI. Importantly,

the rise of big data allied to greater cloud and

computing-processing capabilities boosted

numerous developments in the f‌ield. Nowadays,

AI is not only present as a tool in scientif‌ic

research and industry activities but is also

increasingly in everyday life.

Although there is currently no established

global def‌inition of AI, a recent def‌inition

put forward by the High-Level Expert Group

on AI, set up by the European Commission1,

is presented in Box 7-1. It includes the sub-

disciplines described in Figure 7-1, namely

machine learning (and, within this category,

deep learning and reinforcement learning),

reasoning processes as well as intersections

with robotics f‌ields, for example, sensors.

BOX 7-1 Towards an AI def‌inition

Artif‌icial intelligence (AI) systems are

soware (and possibly also hardware)

systems designed by humans that, given

a complex goal, act in the physical or digital

dimension by perceiving their environment

through data acquisition, interpreting the

collected structured or unstructured data,

reasoning on the knowledge, or processing

the information derived from this data and

deciding the best action(s) to take to achieve

the given goal. AI systems can either use

symbolic rules or learn a numeric model,

and can also adapt their behaviour by

analysing how the environment is af‌fected

by their previous actions.

As a scientif‌ic discipline, AI includes

several approaches and techniques,

such as machine learning (of which deep

learning and reinforcement learning are

specif‌ic examples), machine reasoning

(which includes planning, scheduling,

knowledge representation and reasoning,

search, and optimisation), and robotics

(which includes control, perception,

sensors and actuators, as well as the

integration of all other techniques into

cyber-physical systems) (see Figure 7-1).

Source: European Commission (2019), Report by the High-Level Expert Group on AI set up by the European

Commission