AI in Parliaments: The variable of political systems and the importance of training human resources
Author: Juan de Dios Cincunegui: Director of the International Centre for Parliamentary Studies, Research and Foresight of the School of Government of Universidad Austral. Written on September, 2024
This publication is part of the book “Artificial Intelligence in Legislative Services: Principles for Effective Implementation”. To download the entire book, use the button below:
Introduction.
According to the United Nations, human beings currently live an average of 73.4 years (World Population Prospects:2023). As a result of scientific and technological innovation, among other factors, life expectancy is expected to increase significantly in the future. However, there is a great disparity, since at the top of the ranking there are countries, whose population lives today between 80 and almost 90 years, while others, at the bottom of the list, barely exceed 50 (averages that include both men and women). Until the 19th century, life expectancy was slightly higher than 30 years on average. This, among other factors, had a direct impact on population growth. The planet went from being inhabited by 4 million people twelve thousand years ago to 8.2 billion today. The same planet!
When and how did such a population explosion occur? Just two centuries and twenty years ago. It was in 1800 when the planet reached its first billion inhabitants, amid the Industrial Revolution, a year preceded by great events such as the discovery of vaccines and the development of medicine, which among other achievements allowed a drastically reduced infant mortality rates after their universalisation in 1974.
It was also between the end of the 18th century and the beginning of the 19th century that modern democracies were born, with the United States of America at the forefront, a country that became independent in 1776, adopted and ratified its constitution in 1787 and 1788 respectively, and elected George Washington as its first president in 1789. From then on, one of the most challenging inventions in the history of humanity would be put to the test, which has been gradually perfected until today, coexisting with advances and setbacks.
In the same vein as the demographic phenomenon, the world went from not having a single democratic country in the 18th century to almost 50% of the countries worldwide having democracies of some kind in 2023 (Our World in Data). Progress from this point of view is evident. However, in terms of population, only 7.8% of the world’s inhabitants currently live in full democracies, 37.6% in flawed democracies, and almost 40% in authoritarian countries. The rest live under hybrid regimes (Democracy Index). It is clear, therefore, that there is enormous room for improvements in the political system and the quality of life of the population, but it is important to be aware that the great transformations of humanity described above are relatively recent.
Why do I mention this background?
Because we cannot analyse the AI phenomenon without contextualising it.
The emergence of AI: a powerful and unstoppable revolution.
Hand in hand with great scientific and technological advances, each of humanity’s industrial revolutions has positively and negatively impacted society. The first revolution, mentioned above, was followed by three others, until the current one, the fifth, called “Industry 5.0”, which was just born and is still in the process of consolidation from a theoretical and practical point of view.
The 2nd Industrial Revolution, located between the middle of the 19th century and the beginning of the 20th century, brought with it extraordinary inventions in areas such as agriculture, communications, energy (adding electricity, gas, and oil to steam and coal), chemicals and pharmaceuticals, health, steel, tobacco, transportation, among others, including, for example, automobiles, airplanes, improvements in railroads; cameras, moving images, the phonograph, the radio, the telephone, the electric light bulb, synthetic dyes, plastics, and new technologies for mass production.
The 3rd industrial revolution, which took place in the late 20th century, also brought about immense innovation in the areas of energy and communications, with the development of the nuclear energy and renewable energy industries, as well as the computer and mobile phone industries. These were the days of the Canadian thinker Marshall McLuhan, who coined the expression “global village” to refer to an increasingly interconnected and interdependent world; and who also predicted the coming of the digital revolution, although he was unable to witness the universalisation of the Internet, which took place in the early 1990s.
The 4th industrial revolution, announced by Klaus Schwab very recently, in 2016, and for some still current, is characterised by the intelligent articulation between physical, digital, and biological equipment and systems, including machine learning, big data, biotechnology, cognitive computing, quantum computing, cloud computing, 3D printing, genetic engineering, the Internet of Things, nanotechnology, virtual and augmented reality, advanced robotics, autonomous vehicles, and many other innovations that include artificial intelligence.
One of the most notable features of this 4th industrial revolution is, precisely, the exponential progress of technological innovation, giving rise, for this reason, to the announcement of the arrival of the 5th industrial revolution, focused on achieving a reasonable balance between these new technologies and human beings, putting the latter at the centre of the scene, both from the point of view of the provision of personalised services and in the field of production and collaborative work between robots and humans. “Cobots” stand out as an example, especially in the field of medicine.
AI plays a prominent role in both the 4th and 5th industrial revolutions.
Although awareness of its feasibility dates to the preliminary studies of the British Alan Turing in the middle of the last century, the expression was coined by the American John McCarthy, who was responsible for promoting specialised studies and research aimed at its development just a few years later. Many things have happened since then, especially the exponential increase in the generation and storage of digital data and, in parallel, a huge reduction in computing costs. Big Data, popularised at the beginning of the 21st century, allowed the development of AI to be unleashed to reach unsuspected levels. Progressively, AI began to creep into different areas of our lives, becoming one of the most disruptive factors in the history of humanity. However, its ability to emulate and even improve the intelligence of human beings poses unimaginable challenges.
For example, according to various tests comparing the performance of humans and AI in different fields, the latter has outperformed us. The gap will probably continue to widen over time. I am referring to activities such as handwriting recognition; voice recognition; image recognition; reading comprehension; language comprehension; and predictive reasoning (Artificial Intelligence: Giattino, Mathieu, Samborska and Roser:2023).
ChatGPT, for example, an AI chatbot from Open AI, has become a sort of “digital celebrity.” Its design combines different technologies, including NLP (natural language processing); DL (deep learning); and ML (machine learning). Its performance level is proportional to its growing popularity.
Let’s see below how AI is or could be linked to parliaments.
Logical Framework.
Addressing the incorporation of AI tools in the parliamentary sphere necessarily requires a logical framework.
I pointed out in the introduction to this article that the planet will reach 8.2 billion people by the end of this year and that approximately 50% of the countries on the planet are “democratic.” I also highlighted the differences in life expectancy of the population (on average) depending on the country of residence. Now I would like to relate the AI phenomenon to these variables.
“Parliaments” (in the broadest sense), whatever their name (legislatures, congresses, assemblies, councils, or whatever); political regime (autocracy or democracy); or political system to which they belong (e.g. presidential, parliamentary, semi-presidential or semi-parliamentary), have some common features such as their composition (usually by a group of people chosen through some type of prior selection process); representing the voice of the citizenry (at least “formally”); and the capacity to legislate or advise on legislation. Some parliaments also have the power to review and approve the budget and/or control other branches of government, among other legislative powers.
The Inter-Parliamentary Union (IPU), a sort of “United Nations of Parliaments”, recognises the right of “any Parliament constituted under the laws of a sovereign State, whose population it represents and in the territory of which it functions” to apply for membership. Additionally, “any Parliament constituted in accordance with the fundamental law of a territorial entity whose state vocation is recognised by the United Nations and enjoys permanent observer status at the United Nations, with important additional rights and privileges, may also be a Member of the Inter-Parliamentary Union” (Article 3, paragraphs 1 and 2 of its Statutes, texts in force since March 2023).
The IPU currently comprises 180 national parliaments, including the parliaments of countries considered “autocracies”, whether “full” or “electoral” (in terms of the Democracy Index).
Why do I make this distinction?
Likely, the logic behind the evaluation and probable incorporation of AI into the management of the legislature of a democratic government or an autocratic government has both different and even opposing motivations and objectives.
In 2023, only twenty-four countries around the world were considered “full” democracies. These are, in order of priority, Norway, New Zealand, Iceland, Sweden, Finland, Denmark, Ireland, Switzerland, the Netherlands, Taiwan, Luxembourg, Germany, Canada, Uruguay, Australia, Japan, Costa Rica, the United Kingdom, Austria, Greece, Mauritius, South Korea, France and Spain.
In the same year, the twenty-six countries on the planet with the worst rating, considered in all cases as “authoritarian” regimes, were, in order of priority: Afghanistan, Myanmar, North Korea, Central African Republic, Syria, Turkmenistan, Chad, Democratic Republic of the Congo, Laos, Sudan, Libya, Equatorial Guinea, Tajikistan, Yemen, Iran, Eritrea, Belarus, Saudi Arabia, Uzbekistan, China, Burundi, Gabon, Guinea, Russia, Nicaragua and Venezuela.
It is enough to look at the positions and indicators that make up the Human Development Index prepared by the UNDP, such as life expectancy at birth, average years of schooling, or GDP per capita, to see that the highest levels of development correspond to democratic countries and the lowest to non-democratic countries.
In terms of AI, this distinction will be decisive, since its application will produce different effects depending on the type of society in question.
In democratic countries and regions, AI is generating major, open, and public debates about its positive and negative aspects, the consequence of which will be the adoption of regulations, public policies, and actions in a collaborative manner.
In non-democratic countries and regions, AI, on the other hand, will become an instrument of concentrated power, maximising social control, and further reducing people’s margins of freedom.
Evidence shows that scientific and technological innovation processes have accelerated dramatically and that technologies such as AI will surely continue to grow exponentially. The result will not be innocuous, impacting differently depending on the level of freedom and development of each country or region.
Parliaments and AI.
Parliaments must address AI on at least three fronts:
a). from the point of view of its general understanding and the opportunity, merit, and convenience of legislating on the subject, as well as regarding the feasibility of simultaneously promoting appropriate public policies to achieve certain ends based on a vision and a strategic plan that responds to the interests of society.
b). concerning all types of sectoral policies, including, for example, education, science and technology, production and trade of goods and services, health, security, energy, transport, communications; infrastructure; the environment, migration, tourism, sports, culture, the entertainment industry, human rights, etc., in their aspects of design and planning as well as execution, review, and permanent reformulation; and
c). regarding the functioning of the State itself, including the three branches of government: executive, legislative, and judicial, with special emphasis, from my point of view, on matters relating to electronic/digital government and the improvement of democratic quality.
The global concern is so great that this year the United Nations has decided to dedicate the International Day of Democracy, to be celebrated on September 15, to “the importance of Artificial Intelligence (AI) as a tool for good governance.”
The UN High-Level Advisory Body on Artificial Intelligence (HLAB/AI), composed of 32 experts in different disciplines and with the Office of the Secretary-General’s Envoy on Technology (OSET) as its focal point, has been working collaboratively and in a network with representatives of stakeholders from around the world, including governments, the private sector, and civil society.
It is assumed that the best way to take advantage of the positive aspects of AI and mitigate its risks and uncertainty is through global governance. For this reason, the results of the HLAB/AI work will be part of the discussion on the Global Digital Pact that will take place during the Future Summit. This pact, together with the so-called Future Pact and the Declaration on Future Generations, aims to be anchored both in universally recognised human rights and in the SDGs (Sustainable Development Goals) of the 2030 Agenda, being articulated around five pillars: sustainable development and development financing; international peace and security; science, technology and innovation and digital cooperation; youth and future generations; and the transformation of global governance.
Beyond the spirit of collective construction and the different margins of intervention and influence, the truth is that the nations of the world are divided into two large groups: the “rule makers” (those who make the rules) and the “rule takers” (those who follow them). However, without failing to recognise this reality, national governments, as well as subnational governments and individuals and legal entities in general, will always have a certain margin of freedom to act.
The evolution of AI and its consequences at a global level will undoubtedly impact the three fronts mentioned above, which is why the more conscious, professional, and serious decision-makers, especially legislators, are, the greater the guarantees for citizens. Otherwise, societies whose politicians are not up to the task will suffer by default.
A good example to measure the level of preparation and involvement of national parliaments in the debate on the impact of AI on democracy, human rights, and the rule of law is the draft resolution presented by the co-rapporteurs of Canada and Tanzania, Rempel Garner and N. Lugangira, for debate during the 149th IPU Assembly to be held in Geneva between 13 and 17 October, 2024. The initiative corresponds to the Standing Committee on Democracy and Human Rights.
Related to recommendations for AI management in parliaments, in July 2024, the Westminster Foundation for Democracy (WFD), under the coordination of three editors (Fitsillis, von Lucke, and De Vrieze), published the first “Guidelines for AI in Parliaments”, a joint effort of the aforementioned and nineteen other experts from parliaments and other institutions in 17 countries (Australia, Austria, Argentina, Brazil, Canada, Estonia, Germany, Greece, Hungary, Italy, Israel, Lithuania, Qatar, Spain, Switzerland, the United Kingdom, the United States) and the UNDP.
As the guide argues, AI is already being incorporated into legislative processes, so it is a reality. What is interesting is that its recommendations apply to all types of legislatures, whether local, regional, national, or supranational.
The approach taken is holistic, addressing aspects such as ethics, privacy, security, supervision, system design, and education.
AI undoubtedly offers a transformative opportunity for parliaments. Is that enough? No, it is not. The first question that the leadership of the day must ask itself is: is there a desire for change, for genuine improvement? That is the first hurdle that a parliament must overcome when analysing the AI solutions available on the market.
Experience shows that one of the main reasons why parliaments are increasingly moving away from the efficiency and effectiveness achieved by other human organisations that navigate in competitive environments is the lack of will to improve and change. In many cases, this is even a matter of deliberately produced stagnation. Furthermore, if one compares the quality of management and productivity of parliaments in different parts of the world, one will notice enormous differences. There are good ones, average ones, and bad ones.
The three basic questions we should ask ourselves to formulate a genuine AI strategy in a parliament are: a) what kind of parliament is it? b) what are the reasons why it works the way it does? and c) why do you want to incorporate AI into your processes?
The answers will differ depending on whether the Supreme People’s Assembly (SPA) in North Korea, the Pyidaungsu Hluttaw, Myanmar’s bicameral parliament, the Storting in New Zealand, or the Althing in Iceland.
National parliaments in countries with full democracies will surely be able to adapt to AI in a timely and convenient manner, accompanied by the drive of vibrant societies that demand efficient, effective, and, above all, decent public institutions.
In the introduction of AI into national parliaments in countries with flawed democracies, on the other hand, the risks of both malpractice and deviation appear to be greater.
In the case of national parliaments in countries with authoritarian regimes or full autocracies, the introduction of AI into the political system represents, in theory, a huge risk.
Concerning the first two cases, I suggest that the Leaderships adopt the methodology proposed by The Open Government Institute of the Zeppelin University Friedrichshafen of Germany and the Hellenic Parliament as a starting point to determine both the priorities in terms of AI solutions and the urgencies, mediating an exercise as broad and diversified as possible that includes not only key referents within the parliament but also external referents such as other areas of government, the private sector, and civil society, especially the academic, scientific and technological communities (see in this regard “Initial Surveys on the Prioritisation of Advanced Technologies & AI for Parliaments”, by von Lucke and Fitsilis: 2023).
The rest will come in addition, especially the necessary and unavoidable process for the education and training of human resources in parliament.
The importance of training human resources
The staff of any human organisation is an asset if their conduct and performance contribute to the achievement of its mission. Conversely, if staff are not trained and do not evolve at the pace of progress in education and scientific and technological innovation, they become a liability. Furthermore, organisations that have cutting-edge leadership capable of anticipating the future are the ones that succeed and stand out from the rest.
Parliaments are no exception. Moreover, one of the most common challenges of public management is to adapt to the social and technological progress of the societies to which they belong, which are usually more inclined to adapt quickly and periodically to the updates provided by the market.
An example of this phenomenon is video conferencing platforms. Long before the advent of the COVID-19 pandemic, society had already incorporated not only Skype video calling technology but also Zoom, among other IT solutions. For parliaments, however, adapting to these technologies became a novelty.
Several recommendations can be made to ensure that human resources are highly trained and at the forefront of scientific and technological innovation. This applies not only to permanent and occasionally temporary staff, but also to civil servants, whether career or political, and legislators themselves.
First, a serious parliament committed to the public interest must be constantly alert to any circumstance with social impact. New technologies, in general, are agents of change.
For this reason, the US Congress was truly pioneering when it created the Office of Technology Assessment (OTA) in 1972, inspiring the creation of Parliamentary Offices of Science and Technology in Europe and other parts of the world in the following decades. In parallel, in 1993 the Finnish Parliament created the first Parliamentary Commission on the Future, which would serve as the basis for the creation several decades later of the current legislative commissions on the future in various parts of the world.
Good practices in this area indicate that a suitable parliament not only conducts broad and participatory consultative processes that allow experts from different social dimensions to be heard (government, business, scientific and academic communities, civil society, etc.) but also leads public debate through the processing of information and the demarcation of its scope.
The British Parliament, for example, has conducted several inquiries into AI. For example, the House of Commons Science, Innovation and Technology Committee conducted a public consultation on “Robotics and Artificial Intelligence” in 2016; another on “Artificial Intelligence Governance” in 2022-2023; while the House of Lords Communications and Digital Committee conducted another inquiry on LLMs (Large Language Models) and generative AI earlier this year.
In parallel, the libraries of both Houses of Parliament and the Parliamentary Office of Science and Technology (POST) have produced top-level reports on AI that are freely and publicly accessible to both parliamentary staff and the public.
Publications include analysis of the policy implications of AI; AI development, risks, and regulation; AI policy in the UK; the use of AI in education delivery and assessment; education technology: digital innovation and artificial intelligence in schools; AI and health; AI and employment law; data science skills in the UK workforce; AI in security; AI and the justice system; and automation in military operations. These are supplemented by publications with basic explanations of AI; a bibliography; and even an AI glossary.
A report from April this year indicates that the British government decided not only to update the digital skills of public sector staff but also to reduce the digital divide in society through training of the public. To that end, in 2021, the UK government adopted a National Artificial Intelligence with a 10-year outlook aimed at adapting society to the changes that AI will bring.
This is the right type of ecosystem for the design, development, and implementation of training and coaching programs for legislative staff on AI.
These training sessions must respond to a process of adaptation and genuine change in which the people to be trained feel both responsible and proud to belong to Parliament. In parallel, parliamentary staff, as well as Members of Parliament, should be part of prior consultations for the development of plans for the incorporation of AI in parliament, for example through surveys, planning, and foresight exercises, as described above.
Final reflection
I would like to conclude these reflections by suggesting that we explore, among the most notable contributions that AI can provide to parliaments around the world, especially those with flawed democracies, the greater openness to citizen participation and contribution in the strictest sense of the so-called “online deliberative democracy”, in terms of representation and debate; legislation; parliamentary control; budget and parliamentary diplomacy.
Openness, transparency, and citizen participation are concepts opposed to authoritarianism and the concentration of power.
The incorporation of new technologies such as AI into parliaments must respect certain principles such as accountability and transparency; decision-maker autonomy; ethical and responsible use; human oversight; a prior assessment of their impact on human rights and the need to mitigate risks; safeguarding public trust; ensuring inclusion and respecting diversity; periodically adapting to technological advances; fostering interparliamentary cooperation; achieving genuine public engagement; and respecting applicable legal norms (WFD:2023).
If you live in a full democracy, AI will surely contribute to the approval of laws of higher quality and impact on society; it will improve parliamentary control; it will allow for better planning of the State’s financial management (budget); it will enrich debates within parliament (representation); and it will promote quality parliamentary diplomacy. All of this will surely have an impact on improving the chances of you and your loved ones living longer and enjoying a better quality of life. If, on the other hand, you live in an authoritarian country or under an autocratic regime, AI will surely increase the concentration of power, aggravate the loss of freedoms, persecution, and social control; and your life expectancy and that of your loved ones will be lower than in developed countries.
Ultimately, the success or failure of AI in the parliamentary sphere will depend largely on the robustness or weakness of political systems.