Dear Colleagues,  

There is a phrase in Chinese - 见字如面（jian zi ru mian）- which means “reading this letter feels like meeting with you face to face.” When you open this letter, I know I am writing to and meeting someone who cares deeply about education and believes that technology can play a role in helping alleviate some of the greatest problems in education today. For that, I am deeply in awe and grateful.  

AI among other technologies is developing rapidly, and it will have tremendous potential in fundamentally shaping how we think about, organize, measure, deliver, and manage education. The emergence of Generative AI and LLMs has re-energized generations of prior work in the education technology space and outlined an inspiring vision of how technology will, finally, come to “revolutionize” education.  

Yet, great capability comes with greater responsibility. Just as Professor Mitchell Resnick said, “...generative AI technologies could also be used to support project-based, interest-driven creative learning experiences — but only if we make the explicit, intentional choice to use them in that way."¹ Our understanding of any piece of innovation is never complete without fully considering its potential harmful consequences to all stakeholders, community, and society.

With AIED in particular, its potential harm is as great as its potential benefits, in areas such as ill-represented learning datasets, algorithmic bias and discrimination, lack of model explicability and transparency, and scaling up and reinforcing of the social and educational injustices of the past. Without critically examining, designing, and governing super powerful information technologies such as AI and the social-technical system around which these technologies are applied to the ever-contentious processes, pedagogies, objectives of education, we will be facing considerable ethical and social risks of amplifying undesirable old patterns², with individuals’ livelihoods bearing irreversible consequences.  

Unfortunately, the progress of technological advancement still significantly surpasses our understanding of its harmful effects and how to govern it effectively. If we fail to hold these issues accountable, we will either face two undesired paths³: a technological backlash resulting from the distrusts of educational stakeholders, causing all of us to miss out on the benefit that AIED can bring, or the proliferation of untrustworthy and unethical AIED applications that fail to take into account the risks they pose and can lead to irreparable harm.  

We need to act now. Given its early stage of development of this new generation of AIED, there are plenty of opportunities to bring ethical practices and awareness to all stakeholders (including policymakers, researchers, funders, developers, entrepreneurs, teachers, administrators, parents, and learners) and set up guardrails against harmful consequences of AIED systems. But we cannot afford to waste time, when billions of dollars and trillions of model parameters are deployed in a technological arms race to develop even more advanced technologies that no one is ready to govern. The time is now.  

A team of us - aspiring changemakers and visionaries from Stanford, Cambridge, UPenn - have come together as a non-profit to promote ethical and responsible AI in education. Our organization's name - Learnest ("Learn + Earnest") - speaks to our "sincere and intense conviction" in "learning" as a vehicle for social impact and progress and our determination to harvest the power of AI responsibly for learning. We envision a world in which all learners can benefit from and thrive with AI, and Learnest can play a role in accelerating the realization of this vision.

Fundamentally, we pose a few important strategic and normative questions to the entire community:

• What kinds of educational problems can practically be solved by computers, and by AI?
• What is the relationship between AIED systems and educational ideals, values, goals, and practices?
  • Are they merely continuing old, ineffective, and unjust paradigms more efficiently, or
  • Are they helping pursue a human-centered reform and refinement of the current education system for the betterment of society?
• For those that can be solved by AI, SHOULD we use AI to solve them?
  • Under what circumstances?
  • What are the costs vs. benefits, and for whom?
  • What kind of trade-offs are inevitable?
  • And who gets to decide, govern, and regulate?

As a community, we need to work together to answer these questions and set a clear boundary of what and how AIED systems should be developed and deployed for whom. Working towards this end, Learnest aims to impactful and human-centered innovations in education through training, community building, and research advancement

• Training: Equipping practitioners with practical responsible AI knowledge and skills in education,
• Community Building: Building multi-stakeholder, cross-disciplinary communities to foster understanding and diverse perspectives of AI in education, and
• Research Advancement: Facilitating and disseminating cutting-dge research in the field

Learnest’s first flagship program – the Ethical AI in Education Fellowship - aims to bring together and train the brightest and most innovative minds in this emerging field, as a leverage to accelerate our initiatives and push for change from the bottom up. The fellowship’s mission is to build a movement to inspire a new generation of AIED scholars equipped with ethical principles.

As practitioners and innovators of education, we must always remember that the purpose of education is to carry human civilization forward. We are fortunate to live in a time and space where we stand on the shoulders of giants who come before us with their tremendous wisdom, possess the most advanced technologies that can efficiently and effectively extract these information and knowledge as we need, and pass on these insights to the next generation in unprecedented ways. It is our generation’s mandate to make sure we get this right, so that our future generations can reap the benefits they deserve.

Yours sincerely,
‍
Chenming (Richard) Tang
Founder, Learnest

Original: June 2023; Edited Feb 2025