AI PORTUGAL 2030
Following in the footsteps of mass digitisation, which has already been changing world economy and societal fabric, AI will accelerate and amplify the process. To think ahead and prepare action plans in digital matters is important not only to properly handle the pitfalls but also to take advantage of the emerging opportunities.
The impact on people's life and the challenging pace of change of technology enhance the importance of an AI National Strategy that builds on the analysis of its developments and applications to focus on domain specific questions such as the AI preparedness of Portuguese economic, social, and cultural landscape, the role of government, and how to involve the research community in the entire process. However, the setting upon which the so-called AI revolution will unravel is far larger than an AI Strategy can handle, so INCoDe.2030, in all its five action lines, has to prepare the ground for its sound implementation.
WHAT IS AI?
Artificial Intelligence is a broad term with a large number of formal and informal definitions. For the scope of this initiative we refer to AI as the scientific area and the suite of technologies that use programs and physical devices to mimic advanced facets of human intelligence. AI artefacts can display abilities such as (but not necessarily limited to): autonomy, problem solving, complex planning, negotiation, reasoning, inference, decision making, diagnosis, prediction, monitoring, learning from experience, adaptation to new situations, language understanding and generation, explanation, argumentation, visual/audio recognition, object recognition and the generation of complex artefacts. For this strategy, we will focus on emerging AI technologies.
PORTUGAL IN TECHNOLOGY AND AI
Portugal is showing good results in some innovation indicators (including but not limited to AI), although in many of them we have been typically placed below the average of the European Union*. Portuguese institutions are particularly well positioned in terms of international research collaborations, broadband penetration and product/process innovations in Small and Medium-size Enterprises (SME). Portugal has been relatively successful as an innovation-friendly environment and has an attractive research system.
Human Resources: Numbers from 2017 show that Portugal has a shortage of qualified human resources in advanced technological areas, mostly in terms of higher education (67% of the EU average in 2017) but also in lifelong learning (88.8%) and new PhDs (94%). Employment in knowledge intensive activities is low (57% of EU) but it is slightly above average in fast-growing enterprises (103.2%).
Research: Portuguese research has a high level of international collaboration (185% of the EU average in 2017) participating in the 10% most cited works (82.6%) and in the attraction of foreign PhD students (98.3%).
Innovation: The slice of employment of fast-growing companies in the most innovative sectors has been improving. The R&D expenditure of the business sector has considerably improved since 2015 and represents about 52% of gross expenditure in R&D. SMEs are doing quite well in innovations in the product or the process (158.8%) and in marketing/organisation levels (112%).
Infrastructure: The best indicator of Portugal in the 2018 European Innovation Scoreboard is broadband penetration (200%). Despite us having a low level of Internet usage (below 60% of the households in 2017)** including e-commerce and internet banking with a better performance in the use of social networks.
* “European Innovation Scoreboard 2018” and “Country Report Portugal 2018
Including an In-Depth Review on the prevention and correction of macroeconomic imbalances”, COMMISSION STAFF WORKING DOCUMENT
** Digital economy and digital society statistics at regional level, https://ec.europa.eu/eurostat/statistics-explained/
By 2030, Portugal will have a knowledge intensive labour market with a strong community of forefront companies producing and exporting AI technologies supported by an academia involved in high-level, fundamental and applied research. AI technologies will be easily available to promote the efficiency and quality of all activities, including SMEs, public services and every citizen. The labour force will be highly qualified and Portugal will be at the forefront of AI Education for all. AI will improve the quality of services and the efficiency of processes while guarantying fairness, wellbeing and quality of life.
Our main general objectives to reach by 2030 are the following.
- Added Economic Growth: the added value brought by AI technologies to the economic growth should be significant.
- Scientific Excellence: improve the front-line position in fundamental and applied AI research of the Portuguese Academia (universities, polytechnic schools and research institutions) measured in terms of publication impact, international leaderships, and international collaborations.
- Human Development: Increase dramatically the qualifications of the labour force, in particular the technological qualifications, while promoting inclusion and awareness at all levels of education.
In the process we expect to observe a significant increase in the number and volume of knowledge-intensive AI companies displaying a considerable R&D effort and increased collaboration between academia and companies/public sector. The awareness of AI capabilities and how it can be used to boost business and overall quality of services should spread to the whole of society including SMEs, and the public sector, as well as to citizens in their everyday life.
The growing application of AI must also strengthen societal robustness by building a clear vision of the impacts of AI on democracy, privacy, security, fairness, the labour market, governmental and commercial transparency and equity. Although AI is highly disruptive in all these dimensions it also provides, if made ethical-by-design, a set of powerful tools to actually improve society and democracy.
The strategy will revolve around four main interacting processes.
- The attractiveness of Portugal for knowledge intensive young companies and international production units is high and has the conditions for improvement. These units work in different sectors but have the need for development of specialised AI software and high-tech devices for export in common. The collaboration with academia is growing in two axes: joint research uptake (joint projects and CoLabs) and the qualification pipeline.
- The development of this ecosystem will motivate the increase of the currently developing innovation levels for a vast number of companies and organisations, including start-ups, SMEs, and the Government Sector through business networking and by benefiting from the maturing collaboration platforms with academia. These include AI-on-demand pipelines and Digital Innovation Hubs. Expected outcomes include an increase in the number of patents and the multiplication of innovation-based businesses.
- The research potential in AI and other areas will grow due to the larger share of private investment and because of the added value induced by the challenges brought by innovating companies. Moreover, researchers will gain insight into the future of AI itself as a fundamental scientific field. Expected outcomes are a higher attraction of research talent (and as a consequence of professional talent), a higher impact of scientific publications and an increased ability to join international research networks of excellence. These scientific results will in turn revert back to the productive sectors.
- Academia alone and in collaboration with the Industry will increase its capacity and develop diferent levels of qualification programmes in AI and related areas. Other educational institutions working at different education levels will also be motivated to invest in skilling, reskilling and lifelong learning aiming at tailored qualifications. As an outcome, Portugal will increase the qualification levels of its professionals and increase the level of knowledge-intensive employment.
AI PORTUGAL 2030 > PDF
Areas of specialisation in Portugal with an international impact
Portugal currently has strong players in some areas that may serve as inspiring examples and help drive innovation and research. The following list identifies some of these areas where Portugal will make a specific effort to lead in Europe.
- Natural Language Processing
- Real Time decision making with AI
- AI for Software Development
- AI for Edge-computing
Areas for research and innovation in European and international networks
In the area of AI there is relevant investment of European companies in Portugal in several domains, such as automotive, car multimedia and intermodal transports; information systems; components and services for 5G networks; smart cities and security; banking; and bioeconomy and biorefineries. Some outstanding examples with companies, researchers and joint ventures already in action are referred hereafter.
- Urban transformation through sustainable Cities
- Sustainable energy systems
- Environment and biodiversity: from forests and green economy to marine species and blue economy
- Mobility and autonomous driving
Fundamental research for Future AI
Important research lines are emerging, such as: Transparent AI: giving algorithms the ability to explain their own decisions and provide a high level and adaptive account of their workings to promote fairness and accountability; Emotional AI: algorithms will utilise emotions to achieve better decisions; Autonomous AI: important not only in the automotive sector but also in information systems, cybersecurity, smart cities, industry, etc.; AutoML: intelligent systems that can use machine learning autonomously; Computational Creativity: producing and creating artistic output is nowadays mostly confined to human action. However, there are already good examples, which indicate that machines will have a role to play in the cultural industry as well; Computational Ethics: machines that can acquire, learn, discuss and adapt moral principles through algorithmic procedures.
Public administration and its modernisation
The programme aims to support Research & Development (R&D) activities that contribute to strengthen scientific and technological competencies to deal with the vast amounts of data generated in Public Administration. The goal is to use and combine this administrative data with data from other sources to better inform our public policies and decision-making processes, which should be increasingly supported by evidence and not by intuition. And, ultimately, to change public service provision from a reactive paradigm to an anticipatory paradigm.
Currently 19 R&D projects are being funded through the “FCT's Mobilising programme to foster AI in public administration”: four initial pilot projects aim to support already established and mature partnerships between the R&D community and public administration; 15 additional projects were selected by an independent scientific panel under a specific call aiming to identify and foster new partnerships. These 19 ongoing projects result from the cooperation between scientific institutions and entities of the public administration and are being developed under the Innovation axis of INCoDe.2030. They cover several governance areas such as health, education, transportation and urban mobility, economic activity and land use planning.
Qualification and Specialisation
Qualification and Specialisation, two INCoDE axes, are key to the development of AI driven innovation and economy in Portugal. To qualify human resources at different levels (professional courses, bachelors, masters, post-graduate specialisations, PhD and post-doc) is a challenge for a number of different reasons despite the excellent quality of both our institutions and students.
Even though Portugal is close to the European average in terms of digital competences (15th in the DESI 2017 Index, Digital Economy and Society Index of the European Commission), it needs to reinforce Information and Communication Technology (ICT) competences. This applies to specialists, who need to be able to make the most of the growing availability of jobs in the digital market, but also to the entire human capital, increasing for example internet usage levels, which are still worryingly low.
Portugal offers a training infrastructure as well as the human potential capable of being (re)qualified to meet the demands of employment opportunities that are typical of modern societies such as Portugal. However, this (re)qualification is a demanding task that requires mobilisation and a combination of efforts from different areas of governance and civil society.
Qualification in exploiting AI is, for example, the cornerstone of Industry 4.0. Both Specialisation and Research have to deal with advanced AI techniques and solutions mastering them at a theoretical and technological level, developing and implementing new solutions in many different areas, such as health, space, maritime, industry, agriculture, cities, services and mobility (to name a few).
Inclusion and Education: disseminating generalist knowledge on AI
Digital Inclusion and Education for all is an essential component of AI development. More extensive knowledge and skills will facilitate an adaptation to changes in professions. Opportunities for expanding qualification contents will be integrated in education programmes. The aim is to avoid highly specialised competence in an excessively narrow field.
Digital Inclusion and Education are two main axes of the INCoDe.2030 working towards securing a prominent place in terms of digital skills until the end of the next decade.
- Digital inclusion
New developments and supporting areas in European and international networks
- Advanced computing: supercomputing
- Quantum materials and quantum computing
Facing societal challenges brought by AI: Ethics and safety
AI systems will make important and critical decisions autonomously. Society will demand transparency (the ability to explain the decisions) and auditability (the ability to trace the flow of decisions and actions from humans to algorithm) in order to promote safety and ethical principles, including privacy protection and fairness. We will need best practices to assess AI projects in terms of risks to safety and ethics and mechanisms to detect and prevent misuse of advanced AI techniques. The legal framework will have to be adjusted to determine liability in conflicts with the involvement of AI decision making.
Do you have suggestions on this National Strategy?