How can we measure the contribution of Artificial Intelligence to human development and flourishing?
United Nations (UN) Secretary-General Antonio Guterres called on the UN Security Council on July 18 to emphasize the crucial yet ambivalent role of the latest advances in Artificial Intelligence (AI).
Ambivalent, as Antonio Guterres points out that AI systems have the potential to improve human living conditions and development. The UN, for example, uses such systems to facilitate its humanitarian interventions and ensure cease-fires are respected.
At the same time, the Secretary-General warned of the many dangers exacerbated by simplified access to AI systems. In particular, he cites the ongoing proliferation of content including fake information for the purposes of disinformation and manipulation, emphasizing the change in scale never seen before in our history.
This observation is shared by many other leading players. The governance of these new tools is an extremely thorny issue that many organizations (public and private) are urgently integrating into their strategy.
The Human Technology Foundation (HTF) - with its international, multi-disciplinary network of experts - has also taken up this issue, the importance of which has been emphasized at the highest level by Antonio Guterres.
Unfortunately, research into the opportunities and risks associated with the development and use of AI systems is undermined by a number of factors.
- Firstly, several experts (recognized or self-proclaimed) have immediately diverted attention to the long-term impacts of these tools, by prophesying (with little to no data to support their assumptions) the end of the human species, for example.
- Secondly, to date, there are no clear indicators or methodology for determining whether AI systems as a whole are collectively contributing to progress for humanity, or whether their use is causing a deterioration in our living conditions. AI systems can be used in just about every sector. Consolidating their impact is therefore an extremely arduous task.
- Thirdly, the velocity of development is unprecedented, as is the speed of adoption of these sometimes imperfect AI tools. Antonio Guterres recalls that printing took over 50 years to develop, while ChatGTP reached 100 million users in just two months.
The Human Technology Foundation has therefore set itself the ambitious task of developing an assessment methodology to streamline the ongoing debate and inform decision-makers to enable them to make informed decisions.
To this end, HTF has launched two projects:
- A position paper to detail the opportunities and points of vigilance for new uses enabled by the latest advances in AI (such as Large Language Models and generative AI models). It will be published in September.
- A Well-Being Index to quantitatively measure the contribution or damage of the use of tools such as AI systems on human well-being.
The second task is particularly complex, given the multiple definitions of human well-being. The 2023 edition of the Well-Being Index sets out to list the various definitions (and their underlying measures) of well-being in order to identify the missing elements. To date, our institutions have :
- Purely economic indicators: Gross Domestic Product (GDP) and GDP per capita for instance. There are also indicators such as the Gini coefficient, which measures the distribution of wealth within a population to highlight inequalities within that population.
- Extra-economic indicators focused on the individual: the Human Development Index includes GDP per capita, as well as other equally important data such as life expectancy and education levels.
- Overall extra-economic indicators: pioneering projects such as Gross National Happiness (GNH) include dimensions such as environmental protection and sustainable use of natural resources. It should be stressed, however, that these indicators aggregate a large amount of data: the GNH, for example, is made up of 33 different indicators covering 9 different areas of human life. The United Nations (UN) Sustainable Development Goals (SDGs) have the same ambition, and are increasingly adopted at international level. They include 169 different indicators, making them one of the most exhaustive frameworks for analysis.
Based on the UN's SDGs, the Human Technology Foundation has developed a detailed assessment matrix in its report "Responsible Investment in Technology". This matrix focuses in particular on the impacts linked to the use of digital tools to highlight the clout these tools can have on human well-being.
The difficulty now lies in moving from the qualitative level to the quantitative. How can we measure on a large scale, for example, the changes in our social interactions due to a technological intermediary (instant messaging, social networks)? Furthermore, can an individual's loneliness be attributed solely to his or her use of digital tools?
This is one of the Human Technology Foundation's major ongoing projects, and we invite you to find out more about it in our selection of articles. 🔽
Selection of articles :
- A report by UNCTAD (the United Nations Conference on Trade and Development, responsible for integrating developing countries into the world economy and promoting their development) illustrates the difficulties encountered in measuring the impact of information and communication technologies (ICTs).
- In chapter 2.10 of its report "Measuring the Digital Transformation, A Roadmap for the Future", the Organisation for Economic Co-operation and Development (OECD) highlights the lack of indicators for assessing the impact of digital tools on human well-being.
- The Center for Humane Technology is an advocacy NGO that aims to defend the attention capacities of users of digital tools. In particular, they campaign for more restrictive frameworks to minimize the harms associated with the use of digital tools. They have published a regularly updated compilation of scientific research on the harmful impacts of digital tools on human well-being:
- The Partnership on AI (PAI, a non-profit partnership bringing together technology companies such as Google and Meta, and academic institutions such as Berkeley, as well as associations (the Human Technology Foundation is a member) has published its "Framework for Promoting Workforce Well-being in the AI-integrated Workplace", which brings together the scientific literature on the impact of AI on the workforce. These impacts are classified into 6 categories, and the PAI summarizes for each category the beneficial and detrimental effects of integrating AI systems into the work environment: