As AI advances,
will people advance too?
AI is not just an engineering challenge, it is also a human design problem. For AI to live up to its lofty expectations and benefit humankind, it is important that we not just optimize AI itself and make it more accurate, but that we also understand how people respond to interaction with AI and how we best design that interaction so people and humanity benefit. The Advancing Human-AI Interaction Initiative (AHA!) aims to design and study the interaction between human and artificial intelligence. The goal is to steer AI R&D towards a means to augment and benefit humans. The initiative creates a new research culture around the human experience with and impact of interactive AI, informed by a deep understanding of human needs and behavior. By closely collaborating with AI developers and other stakeholders in a culture of prototyping and experimentation, we aim for this research to have real world impact and move AI deployments in a direction benefiting humanity.
Our Goals
The goals of the AHA! initiative are ambitious and multifaceted. We aim to establish a prominent research field dedicated to the study of human augmentation with AI by using the following three methods of inquiry:
Insight
Prototyping and evaluating novel interfaces and interactions of interactive AI systems that aim to optimize human agency and flourishing for all ages and multiple contexts.
Invent
Performing human subject experiments to understand how choices in the design of interactive AI systems affect user trust, reliance, attachment and perception of the AI system, as well as the resulting behavior of the combined user+AI system in the short and long term.
Inspire
Development and deployment of novel human-AI interactive technologies, experiences and applications that augment people by providing wisdom, wonder and supporting their wellbeing.
A key objective is to bridge the gap between AI developers, the academic research community and targeted AI users by fostering an environment of joint design, experimentation and knowledge sharing that benefits all. We are committed to creating tangible, open AI technologies and solutions —tools, models, prototypes, datasets, best practices, open-source projects and even startups and even spinoffs— ensuring the initiative has practical and meaningful impact.
Pervasive AI offers unprecedented possibilities:
- Health: personalized monitoring, diagnosis, interventions for physical and mental health
- Learning: Personalized, always present tutor for situated learning
- Work: Amplifying creativity, performance, discovery, decision making and more
- Elderly: supporting healthy independent aging
- Entertainment: Immersive, personalized experiences and interactive storytelling
- Social Justice: Data-driven policy making, bias detection, and equitable resource allocation
But also unprecedented challenges:
- Overreliance
- Misplaced trust
- Loss of human agency
- Deskilling
- Loss of understanding
- Misinformation
- Manipulation
- Loss of control
- Privacy erosion
- Unhealthy attachment
- Dehumanization
Why Focus on Human-AI Interaction?
We are in a period of rapid development in AI, with most of that development happening in industry rather than academia. Developers of foundational AI models and applications are in a race to reach AGI spending most of their time, energy and attention on engineering challenges and optimization of the models’ behavior. While they are very engaged in issues such as improving accuracy, efficiency and reducing bias, they devote little to no attention to understanding how people respond to interacting with seemingly intelligent systems and how to best design interfaces and applications to maximize desired outcomes of the use of these systems in a human context.
Experts have warned that having AI at one’s fingertips may lead to overreliance on AI, loss of skills by people, and unhealthy social attachments to companion bots. In addition, many fear AI may lead to negative mental health outcomes and even societal revolt as people gradually are surpassed by AI in the highest intellectual endeavors including creativity, reasoning and problem solving. The AHA! initiative aims to understand people’s responses to AI in the shorter and longer term, with the aim of designing AI such that people have a fulfilling role to play, and remain in control of our creations. While several non-profit organizations have been created to study the ethics of AI, AI safety issues, and to develop open AI models and resources, to our knowledge there is no significant effort dedicated to understanding the impact of AI on people’s beliefs, behavior and mental wellbeing.
Featured Research
How might we design AI to…
Enhance learning motivation
How can AI-generated characters be effectively designed and implemented to enhance learning motivation in educational settings?
Encourage long-term thinking
In what ways can AI Digital Twins be utilized to encourage and develop long-term thinking in individuals?
Promote critical thinking
What strategies can be employed to leverage Wearable AI technology in promoting critical thinking skills among users?
How might we design AI to prevent …
False Memories
What methods can be developed to identify and mitigate the occurrence of AI-induced false memories in human cognition?
Deceptive Explanation
How can individuals recognize and resist AI-generated deceptive explanations in various contexts?
Over-reliance
What interventions can be implemented to prevent over-reliance on AI systems while maintaining their benefits in decision-making processes?
Activities of the AHA Initiative
Through a series of targeted activities, the initiative aims to foster collaboration, innovation, and understanding in this rapidly evolving field.
Research
Development and deployment of novel human-AI interactive technologies, experiences and applications that augment people by providing wisdom, wonder and supporting their wellbeing.
Annual Conference
This event will bring together leading experts, researchers, and industry professionals to discuss the latest advancements and challenges in human-AI interaction. Initiative partners can expect to gain exposure to cutting-edge ideas, connect with thought leaders in the field, and learn best practices grounded in research results.
Focused Workshops
These intimate, hands-on sessions will delve into specific areas of AI and human interaction, providing participants with practical skills and deeper insights. Partners will benefit from direct access to expert-led training and collaborative opportunities.
Newsletter
Our regular newsletter will keep the community abreast about the latest news, research findings, and events related to human-AI interaction. Partners will receive curated content that enhances their knowledge and keeps them abreast of industry trends.
Speaker Series
The initiative will sponsor a hybrid speaker and interview series that will feature some of the key researchers and thought leaders in this area
Outreach
We are committed to engaging with a wide range of stakeholders, including academia, industry, government and the general public. Partners will be instrumental in shaping public discourse and policy through these outreach activities.
People and partners
We envision this as an initiative for a highly interdisciplinary group of AI enthusiasts and experts, which we hope will lead the initiative to critical research and application findings. We plan to collaborate with faculty across the institute as well as external partners committed to understanding, developing and deploying human AI interactive systems.
Join Us
We envision this as an initiative for a highly interdisciplinary group of AI enthusiasts and experts, which we hope will lead the initiative to critical research and application findings. We plan to collaborate with faculty across the institute as well as external partners committed to understanding, developing and deploying human AI interactive systems.