ÎÞÂëרÇø

Addressing Accuracy, Misinformation, and User Understanding in Synthesized Search Results Overview

Many major technology companies ( , ) have incorporated AIpowered overviews that synthesize search results into concise summaries. Some might not understand what an AI overview is, how it differs from traditional search results, what sources it uses, or what risks are involved when relying on it. This can create a in disclosure and digital literacy.

The Challenge

In this year’s IDeaS technology policy challenge, we invite participants to propose usercentered disclosure requirements for AI overviews. What do users think about AI overviews, and how can design and disclosure improvements enhance their evaluation? We anticipate submissions to address the following:

 Overview of the disclosure problem – Including a review of what users currently understand about AI overviews, how platforms presently disclose automated content, and what types of disclosure standards exist in other fields.

 General stakeholder analysis – What do platforms, publishers, users, and civil society want?

 Proposed disclosure framework – Develop a standard or tiered disclosure system based on the topic or “risk-level” of the query, including a detailed description of which queries require specific disclosures, how disclosures are triggered, and how errors are identified and addressed.

 Proposed algorithmic transparency standard – Determine how much and which parts of the algorithm should be shared with the public or researchers, and at what level of detail. Consider which sources to include and which to exclude, and how to handle conflicts between them.

 Implementation Strategy and Impact – How would your standards be put into practice? Consider self-regulation versus government regulation, technical feasibility, economic costs of implementation, and enforcement challenges.

 Evaluation and Conclusion – How should success be measured and tracked? How could these disclosure requirements be modified in the future if necessary?

Example AI Overviews

 Examples to look at (AIOverviews-Examples.csv file)

Rules and Eligibility 

Anyone is eligible to apply. Participants may work individually or in teams.

 Each submission must have an individual registered and attending the conference.

 Entries must represent original work that has not been previously published or submitted to other challenges.

What to Submit

You need to submit two items:

 Policy-Oriented Technical Report: A report addressing the challenge problem, approximately 15 pages (including figures and references), with an executive summary at the beginning. Be sure to include the names, email addresses, and institutions of all team members.

 Promotion/teaser Slide: This is a single PowerPoint slide. Its purpose is to encourage people to visit your poster. It will also be shared online and used to promote your submission. The slide should include the project title, the names of all team members, and their affiliations.

Poster: You will also need to bring a poster summarizing your Challenge solution for presentation at the conference. Posters must be 3’ x 4’ and brought to the conference, as the conference is unable to print posters.

Evaluation

Entries will be evaluated by a panel and community voting during the conference poster session. The winning team of the challenge problem will be invited to submit a full, evidence-based position paper that addresses interdisciplinary issues related to this topic for publication in the Journal of Computational and Mathematical Organization Theory.

Submission Dates

 Challenge Response Submission: August 21st, 2026

 Author Notification: September 4th, 2026

 Final Files Due: October 1st, 2026

How to submit your entry

Submissions will be accepted at the conference installation on

Please email centerforideas@andrew.cmu.edu  with any questions.

Click here for the PDF of the Challege Problem