Many concerns have arisen about AI, related to the lack of knowledge of the subject. Indeed, the term artificial intelligence refers to “the set of techniques that allow a machine to simulate human intelligence, in particular to learn, predict, make decisions and perceive the surrounding world. »
AI implementations are based on data and algorithms, i.e. sequences of instructions to accomplish a complex action.
AI has become ubiquitous and is shaping our lives more than ever.
Two statements, one Canadian and one European, will be presented under two articles. Their purpose is to set out principles for the development of AI.
The Declaration of Montreal, created in 2018 by 500 participants, aims to make AI accessible to the greatest number of people by setting out 10 principles and ethical values in the interest of everyone. These principles are guidelines for a digital transition and a good use of artificial intelligence while preserving from the risks it can generate. They provide an “ethical compass to guide the development of artificial intelligence towards morally and socially desirable ends. »
This declaration is the result of a collective project involving citizens, experts and public officials. It is the responsibility of the various actors and decision-makers to ensure that the deployment of artificial intelligence is compatible with the development of human beings, which is expressed through their emotional, moral and intellectual capacities. This declaration provides a moral framework for analysing problematic social situations and responding to them through ethical and responsible practices.
This reflection is also based on concerns that could be generated by the use of machines and robots, particularly in the educational and medical fields. The aim is to identify principles of good practice in both the public and private domains.
AI, which has grown exponentially since the 1950s, requires an evolutionary approach. As a major scientific and technological advance, it can bring considerable improvements to our living conditions, the environment and the climate.
These major benefits can nonetheless lead to major upheavals if processes such as the infringement of individual rights and the creation of inequalities are not brought under control.
The Declaration of Montreal is based in its construction on 7 fundamental values: autonomy, justice, well-being, privacy, democracy, knowledge and responsibility.
Thus, 10 principles are derived from it, allowing to increase the confidence of individuals in artificially intelligent systems. These must be modulated according to the political, cultural, social and legal contexts.
WELFARE PRINCIPLE :
“The development and use of Artificial Intelligence Systems (AIS) should increase the well-being of all sentient beings. “In other words, the aim would be to improve the living and working conditions of individuals by enabling them to exercise their physical and intellectual capacities without generating a feeling of unease.
PRINCIPLE OF RESPECT FOR AUTONOMY :
“AIS must be developed and used with respect for the autonomy of individuals and with the aim of increasing the control of individuals over their lives and their environment. “AIS must therefore enable individuals to achieve their life goals by providing access to different types of knowledge in order to develop their skills and critical thinking. However, AIS should not be used to influence (propaganda, dominant ways of thinking) or monitor individuals and should not lead to confusion between humans and these systems.
PRINCIPLE OF INTIMACY AND PRIVACY :
“Privacy and intimacy must be protected from the intrusion of SIA and Personal Data Acquisition and Archiving Systems (PDAS). “AIS must allow digital disconnection and ownership of personal data must be respected. Also, individuals must be able to preserve their freedom of thought and emotion as well as their freedom to make decisions.
PRINCIPLE OF SOLIDARITY :
“The development of SIA must be compatible with maintaining links of solidarity between people and generations. “AIS must promote human relations, combat social isolation and improve risk management. Concerning the professional aspect, AIS must promote collaborative work with humans.
PRINCIPLE OF DEMOCRATIC PARTICIPATION :
“AIS must meet the criteria of intelligibility, justifiability and accessibility and must be open to democratic scrutiny, debate and control. “All users must be able to identify whether they are talking to an AIS or to a person and to know by whom decisions affecting them are made.
EQUITY PRINCIPLE :
“The development and use of AIS must contribute to the achievement of a just and equitable society. “In other words, AIS must help to combat discrimination, inequality and social precariousness. Also, access to digital resources must be accessible to everyone.
PRINCIPLE OF INCLUSIVENESS OF DIVERSITY :
“The development and use of AIS must be compatible with the maintenance of social and cultural diversity and must not restrict the range of life choices and personal experiences. “AIS must not reduce individuals to personal identities fixed by data processing. For each service, the offer must remain diversified in order to respect individual freedoms.
PRECAUTIONARY PRINCIPLE :
“All those involved in the development of AIS must exercise caution by anticipating as far as possible the adverse consequences of the use of AIS and by taking appropriate measures to avoid them. “AIS must comply with reliability, security and integrity criteria and restrict the use of algorithms for sensitive areas. Errors concerning AIS and SAADs must be publicly shared by the public institutions and companies concerned.
PRINCIPLE OF ACCOUNTABILITY :
“The development and use of AIS must not contribute to a disempowerment of human beings when a decision has to be taken (crime, misdemeanour…). “In the event of a problem, if a harm has been inflicted by an AIS recognised as reliable, the persons involved in its development or use cannot be held responsible.
PRINCIPLE OF SUSTAINABLE DEVELOPMENT :
“The development and use of AIS must be carried out in such a way as to ensure a strong ecological sustainability of the planet. “Any responsible actor in the development of AIS must support environmentally responsible development projects in order to preserve natural resources, reduce pollution and waste and increase energy efficiency by minimising greenhouse gas (GHG) emissions. Thus, AIS must minimize impacts on ecosystems and biodiversity at all stages of their life cycle.
It is important to develop education around the understanding of the digital ecosystem, including AIS, in order to forge consent and to be in a position to verify the relevance and parameters used in decision making. This education will also enable individuals to acquire various skills to navigate information, protect tools and personal data, share content, etc.
Learning is therefore essential in order to understand one’s environment, to be able to act, criticize and contest certain situations. This corresponds to digital literacy, in other words, “the ability of an individual to manage, understand, integrate, communicate, evaluate, create and access information safely and appropriately using digital tools and networked technologies to participate in economic and social life. »
This learning allows for responsible environmental practices and raises awareness of the energy costs of AIS. The population could thus learn how to repair objects rather than throwing them away, in order to limit digital waste.
However, this digital literacy faces inequalities in access to and contribution to information and knowledge.
Moreover, the predictive calculation proposed by these systems seems to accentuate these inequalities and produce discrimination. For example, Amazon treated their customers differently according to different criteria, including their place of residence, and did not offer delivery service in neighbourhoods where the population was predominantly African-American.
AI has a significant potential for environmental management, it can, for example, allow the automatic monitoring of territories rich in biodiversity in order to preserve our planet.
There are also information schemes on the ecological footprint of products, such as the Type 1 ecolabel (ISO 14024), which guarantees consumers information on the environmental performance of the product and its life cycle. For more than 20 years, ecodesign approaches that integrate social and ecological criteria right from the design and development phase of products and services have been increasing.
With AI, new predictive tools for ecological transition have emerged. Digital technologies already offer information on ecology, short food circuits and teleworking… In the same logic, AIS offer solutions labelled as “AI for the planet” to fight against the ecological crisis. This labeling makes it possible to develop tools in favor of the environment such as :
- A predictive knowledge tool on environmental and social issues such as climate change, biodiversity and agricultural productivity.
- A predictive optimization tool for energy in buildings.
- A tool for predictive regulation of the environmental effects of economic actors.
Thus, the processing of a large amount of data by AI can lead to a better understanding and modeling of the terrestrial ecosystem. To illustrate, organic and sustainable agriculture can be impacted by extreme climate change. If AI could predict these changes, it could help preserve agricultural land.
Although the use of AI has great potential, it must be used rigorously in order to make a significant contribution to the environment. There is a risk of “pathway dependency” – a “social mechanism by which technological, organizational or institutional decisions that were considered rational at one time but have become suboptimal today. »
The AI may have some flaws. For example, if an AIS project enabled urban transport to be optimised with a traffic fluidisation tool, would users decide to increase the distance between their place of residence and their place of work and thus pollute more by participating in urban sprawl? This would give rise to a rebound effect, in other words a “mechanism by which greater energy efficiency or better environmental performance of goods, equipment and services leads to a more than proportional increase in their consumption. “The C40 could create a community for sharing knowledge and experience on climate change and challenges with environmental potential. Open data” could also allow various actors to develop innovative projects on these environmental challenges with a limited cost on data.
Many institutes advocate a “data culture” at the service of ecology with knowledge inputs so that everyone can read, create, use and communicate data.
Thus, by all these points, the digital transition today presents one of the most important transformations of our societies.
To find the second part of the article, please refer to this link.