AI Policy Corner: Is “AI for Good” Overused?

✍️By Priyanka Paradkar

Priyanka is an Undergraduate Student in Computer Information Technology with a minor in Business Management, and an Undergraduate Affiliate at the Governance and Responsible AI Lab (GRAIL), Purdue University.

📌 Editor’s Note: This article is part of our AI Policy Corner series, a collaboration between the Montreal AI Ethics Institute (MAIEI) and the Governance and Responsible AI Lab (GRAIL) at Purdue University. The series provides concise insights into critical AI policy developments from the local to international levels, helping our readers stay informed about the evolving landscape of AI governance. This piece analyzes AI for social good (AI4SG) as a newly adopted research field, and compares China’s and the EU’s approaches to AI for Good.

Overview

In recent years, with the way AI is taking over multiple job scenes, being used in courtrooms and trials, and automation, understanding how to deploy socially good AI initiatives is increasingly important. Thus, AI for social good (AI4SG) has been newly adopted as a research field that focuses on addressing critical social, environmental, and public health challenges by using AI. The UN has outlined 17 Goals that this framework must be in accordance with. As such, the United Nations lists some of these goals to be: No Poverty, Zero Hunger, Reduced Inequalities, and Industry Innovation and Infrastructure.

While this AI4SG field aims to see AI technologies in a positive light and address injustices, many organizations still suggest the need for scrutiny and critical evaluation. Some nations and their frameworks define AI4SG loosely, while others anchor it to enforceable standards and measurable outcomes.

This raises the following questions when thinking about nations implementing socially good AI initiatives: How are countries going to standardize what constitutes an AI4SG initiative? Do different countries’ frameworks align with the same definition for AI4SG?

Differences: A Comparison of China’s and the EU’s Approaches to AI for Good

China’s National Technical Committee on Cybersecurity published an AI Safety Governance Framework, documenting an approach to developing AI for good with the goal of defusing AI Safety risks. This framework consistently mentions the term “AI for good”, which is a recurring theme in the document. China connects the term “AI for good” and governance to national sovereignty and social stability.

China’s governing principles in this framework emphasize the safeguard of national sovereignty, security, and development interests. This idea is listed even before any mention of individual rights, which demonstrates China’s interpretation of AI for good. China makes it a point to define AI4SG in terms of what’s good for the state and social order, while other frameworks and countries adopt a different stance.

For instance, the EU’s AI Act, adopted in 2024, is the world’s first comprehensive binding AI law. The four-point summary outlined in here ties AI for good to individual rights and democratic values, signaling a stark difference in how the EU and China define AI4SG.

Main Takeaways

AI4SG has taken off as a new field to promote the design, development, and deployment of AI initiatives that address social and environmental injustices. However, with the way AI technologies are rapidly developing, organizations and nations have different understandings of what this term means, and as such, can use this loosely defined term to their advantage to promote their technologies.

Therefore, it is just as important to understand the critiques and shortcomings of the term AI4SG, especially with the way it is being used in multiple frameworks and organizations. China’s framework, for example, invokes “AI for good” by framing national sovereignty and supply chain self-reliance as safety priorities. This framing diverges from the EU’s focus, which emphasizes democratic values and individual rights.

This division illustrates how the discrepancy in a common, standardized definition allows “AI4SG” to be shaped around national and security interests instead of the greater social good, which highlights the potential false promises of this initiative.

Further Readings

Image link: https://www.itu.int/initiatives/digital-unga2025/programme/full-schedule/