For the past 20 years, NCTDR has hosted a series of ODR Forums in locations around the world. For 2021, the Forum was held virtually, with live presentation over a web video platform, and recorded presentations available to participants. A full recording of the sessions can be found through http://odr.info/2021-virtual-odr-forum-now-live/. The following items are narrative notes from some of the presentations:

• David Allen Larson – ODR Accessibility

• Noam Ebner – Human Touch

• Jan Martinez & Amy Schmitz – ODR and Innovation

• Frank Fowlie – Online Sport Dispute Resolution

• Larry Bridgesmith – AI Introductory Notes

• Julie Sobowale – AI and Systemic Bias

• Clare Fowler – DEODRISE

• Michael Wolf – ODR 2.0 System Design

• Chris Draper – Algorithmic ODR

• Zbynek Loebl – Open ODR

This is a written representation of the presentation given on 29 October 2019, at 3:20 pm Eastern at the NCSC ODR2019 Summit held at the Colonial Williamsburg Lodge in Williamsburg, VA.

There is significant concern in the access to justice community that expanding current count-based online dispute resolution (ODR) efforts will further exacerbate the systemic inequities present in the American justice system. This well-founded fear stems from the fact that current ODR tools typically calibrate artificial intelligence (AI) algorithms with past outcomes so that any future cases are consistently analysed and filtered in a manner that produces similar results. As courts consider ODR tools for more complicated cases that often require mediation, there is significant disagreement on whether it is possible to create an AI mediator and how that could be achieved. This article argues that an effective AI mediator could be created if its design focuses not on the outcomes achieved by the mediation but on the manner of the communication prompts used by the AI mediator.