How ProSocial AI Can Catalyze Quality of Life and Positive Social Change at Scale: Double Literacy and the ProSocial AI Index
This conceptual article introduces the ProSocial AI-Quality of Life Catalyst Model (PAI-QOL), a multilevel framework defining ProSocial AI as technology designed to generate fair, multidimensional quality-of-life gains through four conversion pathways and enabling conditions like Double Literacy, ultimately offering a disciplined approach to expanding human agency and social value within planetary limits.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Artificial intelligence is no longer just a tool for solving math problems or recognizing faces; it has become the invisible infrastructure of daily life. It decides what news we see, how our work is organized, which medical advice we receive, and how we connect with one another. Because these systems are woven into the fabric of society, their success cannot be measured simply by how fast they run or how accurately they predict the weather. Instead, the true measure lies in how they affect the quality of our lives. This concept, known as quality of life, looks beyond basic happiness to include our ability to make choices, the strength of our relationships, the fairness of our institutions, and the health of the planet we share. When we ask whether a new technology is truly good, we must ask if it helps people live with dignity, if it strengthens our communities, and if it leaves the world better for those who come after us.
A new conceptual framework proposed by researcher Cornelia Walther of Sunway University offers a way to answer these questions. The paper introduces a model called the ProSocial AI-Quality of Life Catalyst Model. Rather than treating artificial intelligence as a force that automatically improves or harms society, the author suggests viewing it as a catalyst. In chemistry, a catalyst is a substance that speeds up a reaction without being consumed by it; its effect depends entirely on the materials and environment around it. Similarly, artificial intelligence can accelerate positive changes like learning and care, or it can accelerate negative trends like dependency and environmental strain. The direction of this acceleration depends on how the technology is designed, who controls it, and what conditions surround its use.
The core of the paper defines a specific kind of artificial intelligence called "ProSocial AI." This is not just a system that follows ethical rules or claims to be helpful. It is a system deliberately designed to produce clear, measurable improvements in people's lives that are shared fairly among different groups. It must also protect the ecological systems that support life and preserve the ability of future generations to solve their own problems. To achieve this, the author argues that we must move beyond vague promises of "responsible" technology and focus on three specific commitments: being pro-people, pro-planet, and pro-potential. Being pro-people means prioritizing human well-being, dignity, and the freedom to make choices. Being pro-planet means ensuring that the technology does not damage the environment or shift environmental burdens onto vulnerable communities. Being pro-potential means making sure that using the technology does not weaken our human skills or reduce our ability to think and act independently in the future.
The paper outlines four specific pathways through which artificial intelligence can actually improve quality of life. The first is the agency and capability pathway. Here, the goal is to help people understand their options and act on them without making them dependent on the machine. For example, a tool that helps a student learn a concept is valuable, but only if it does not replace the student's own ability to think through the problem. The second pathway is relational and social capital. This focuses on how technology affects our connections with others. A system that helps a lonely person find a community is beneficial, but a system that replaces human conversation with a simulated one risks weakening real social bonds. The third pathway involves institutional and public value. This looks at how AI affects trust in organizations like hospitals, schools, and governments. A system that speeds up service is good only if people can still understand the decisions being made and have a way to challenge them if they are wrong. The final pathway is regenerative and intergenerational. This considers the long-term impact on the environment and future generations, asking whether the benefits of the technology outweigh the resources it consumes and the waste it creates.
For these positive outcomes to happen, the paper identifies four essential conditions that must be in place. The first is what the author calls "Double Literacy." This means that people need to understand both their own human needs and emotions, and also how the artificial intelligence system works, including its limitations and biases. Without this dual understanding, people may blindly trust the machine or fail to use it effectively. The second condition is meaningful participation. The people who will be affected by the technology must have a real say in how it is designed and used, rather than just being treated as data points. The third condition is contestability, which means that if a decision made by the system is wrong or unfair, there must be a clear and accessible way for a person to challenge it and get a human review. The final condition is proportionality. This asks whether the use of powerful, resource-heavy artificial intelligence is actually necessary for the problem at hand, or if a simpler solution would work just as well without the heavy environmental cost.
To help evaluate these complex interactions, the paper proposes a new way of looking at the effects of technology, described as a four-by-four lens. This lens asks us to consider four different aspects of human experience—aspirations, emotions, thoughts, and physical sensations—across four different levels of society: the individual, the community, the nation, and the planet. For instance, when looking at an educational tool, we should not just ask if it improves test scores. We must also ask how it affects a student's anxiety, whether it helps them think more deeply, how it changes the relationship between teachers and students, and what the environmental cost of running the system is. This approach ensures that we do not miss hidden harms or unintended consequences that might appear in one area while improvements show up in another.
The author also warns against common ways these systems can fail. One risk is "prosocial washing," where companies claim their technology is good for society without providing real evidence to back it up. Another risk is paternalistic optimization, where the system decides what is best for a person without their input, effectively removing their freedom to choose. There is also the danger of substituting human relationships with synthetic ones, or shifting the hidden costs of the technology onto workers or distant communities. To prevent these failures, the paper suggests that institutions should start by identifying a real human need and then determine if artificial intelligence is the right tool to solve it, rather than starting with the technology and looking for a problem to fit it.
Ultimately, this framework suggests that the future of artificial intelligence depends on human choices. The technology itself does not have a moral direction; it amplifies the intentions and structures of the people who build and use it. By focusing on demonstrable improvements in quality of life, protecting human agency, and ensuring that the benefits are shared fairly, society can guide artificial intelligence toward a future that supports both people and the planet. The paper concludes that while the path is demanding, it offers a clear way to distinguish between technology that merely exists and technology that truly serves the common good.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.