Collaborative AI Conversations: Meredith Ellison, Association for the Advancement of Artificial Intelligence
The rapid growth of artificial intelligence has changed the way we think about research and policy. In response to these changes in the AI space, the Association for the Advancement of Artificial Intelligence (AAAI)—the premier scientific society dedicated to advancing the scientific understanding of thought and intelligent behavior in machines—proactively worked with researchers and policymakers to provide some guidance on navigating the future of artificial intelligence.
Meredith Ellison, DBA, CAE, Executive Director of AAAI, along with other individuals on the Presidential Panel on the Future of AI Research, was part of an effort to identify both the opportunities and challenges of emerging trends in AI. The Presidential Panel was developed to advance the organization’s mission of promoting responsible AI research as well as to increase public awareness and understanding of AI.
The Panel published a report featuring 17 chapters—each covering a topic related to AI research (see below for a detailed list)—that tracks its history, current trends, and open challenges. The study was conducted by 25 AI researchers and supported by 15 additional contributors and 475 respondents to a community survey.
In this conversation with Association Adviser, Meredith provides insight into how the group identified and prioritized the most significant issues—from the promotion of ethical AI development to the use of AI in scientific discovery and autonomous research.
Association Adviser: AI research is evolving quickly across capabilities, methodologies, working environments, and the broader research community. Why was this the right moment for AAAI to convene the Presidential Panel on the Future of AI Research?
Meredith Ellison: Because AI is again a hot topic in the media, the timing felt appropriate to highlight that there is still a great deal of research and work to be done. I think the report also highlights how AI can and is being used across a multitude of research areas and highlights some of the positive impacts it can have.
AA: The panel addresses 17 topics related to current trends and open challenges in AI research. How did AAAI determine the most pressing challenges and trends?
ME: The presidential panel report was spearheaded by Francesca Rossi during her term as AAAI President. Francesca and I collaborated on potential topics and panel members and working with that group, identified key areas of focus as well as additional experts to contribute to the report.
AA: As the work came together, what were some realizations about AI research?
ME: I think the report highlights the vast nature of research in AI and that there is more to be done. I also think that the report itself is not finished. There are chapters not yet written, and we plan to add to the report in the coming months.
AA: The report reflects input from researchers and a broader community survey. Why was it important for this conversation to include such a range of perspectives?
ME: The AI community is vast and covers a multitude of research areas. The same can be said for the makeup of the people doing the research and work so it was important to include as much of the community as possible in the report, not just the smaller cadre of experts.
AA: The report was intentionally written in a non-technical way to reach audiences beyond AI researchers, including policymakers, funding agencies, the media, and the public. Why was accessibility such an important part of this effort?
ME: While much of the work that is produced by authors and published by AAAI is highly technical, it was important to create a report that could be read and used in a more mainstream environment. The goal was to present the work in an easy-to-digest manner, including hosting Zooms on many of the topics, to establish a broader reach and understanding of where AI research is headed.
AA: AAAI-26 in Singapore brought together 10,000 attendees this year. At a time when AI research is evolving so quickly, what can large-scale, in-person events make possible that other formats cannot?
ME: One of the core goals of bringing together the larger research community is for authors to advance their own work through discussions with others. These in-depth conversations just don’t take place virtually, so convening one of the largest AI conferences for those interactions to take place is vital to advancing the research.
AA: For organizations trying to support thoughtful conversations around emerging AI issues, what do you think matters most: access to expertise, clarity of communication, ongoing dialogue, or something else?
ME: I think understanding the goal or purpose of using AI is one of the best places to start. AI is such a broad field and terminology, so the approach to using it should be different for organizations. I would hope that at this point most organizations have already defined how and why they want to use specific AI tools and are keeping up to date on new advancements. There should also be alignment with any governing bodies, boards, etc. to ensure a clear vision and purpose is readily understood by all involved.
AA: What advice would you offer to organizations that want to stay informed and responsible as they follow developments in AI?
ME: I think that this depends on how in the know people want to be. Personally, I get a daily digest from “The Rundown AI” and “The Rundown Robotics” to stay current on new releases, what’s in the pipeline, and other items of note. I am also aware of and see where the core research is heading because of the tens of thousands of paper submissions to AAAI’s conferences.
17 Topics in the AAAI Presidential Panel Report
- AI Reasoning—How AI systems draw conclusions, make decisions, and solve problems in ways that resemble and complement human reasoning.
- AI Factuality and Trustworthiness—Ensuring AI systems produce accurate, reliable information and avoid hallucinations or misleading outputs.
- AI Agents—Autonomous systems capable of taking actions, pursuing goals, and interacting with digital or physical environments.
- AI Evaluation—Methods for measuring AI performance, safety, and effectiveness.
- AI Ethics and Safety—Frameworks and safeguards that guide responsible development and prevent harmful consequences.
- Embodied AI—AI systems integrated into physical bodies—robots or devices—that perceive, move, and act in the world.
- AI and Cognitive Science—Research exploring how AI can model and reveal insights about human cognition and learning.
- Hardware and AI—The physical infrastructure needed to support increasingly powerful AI models.
- AI for Social Good—Applications of AI that address societal challenges such as health, education, equity, and disaster response.
- AI and Sustainability—How AI can both support environmental goals and reduce its own energy and resource footprint.
- AI for Scientific Discovery—Using AI to accelerate breakthroughs in fields like biology, physics, chemistry, and materials science.
- Artificial General Intelligence (AGI)—Research into AI systems with human‑level cognitive abilities across many tasks.
- AI Perception vs. Reality—Examining and closing the gap between public narratives about AI and what the technology can actually do.
- Diversity of AI Research Approaches—Encouraging multiple methodologies to advance the field.
- Research Beyond the AI Research Community—Integrating insights from other professions.
- Role of Academia—How universities contribute to foundational research.
- Geopolitical Aspects and Implications of AI—How AI shapes global competition, national security, and international policy.
Photo courtesy of BOY ANTHONY/Shutterstock.com.