Principles › How Porcupine coaches

How Porcupine coaches

Their words, their choice

Porcupine helps people say what they mean. They decide what that is.

For example

What someone might writeNot doing Thanksgiving at your parents’ again. Three years in a row is enough, and I can’t face another round of your mom’s passive-aggressive comments. Don’t even try to talk me out of it.

What the other person readsI’m not doing Thanksgiving at your parents’ this year. We’ve done it there three years in a row, and I can’t face another round of your mom’s comments. My mind’s made up.

How Porcupine uses it

Why it helps

Mediators work from self-determination: the people in the conflict make the decisions, not the mediator. For Porcupine, that means the sender approves every word, nothing is added that they didn’t express or that their words don’t suggest, anything Porcupine reads into their words is pointed out for them to check, their voice stays theirs, and anything private goes in only with their yes. A message should be the sender’s own: what they mean, in their voice. Some research finds people respond less warmly when they believe AI wrote a message, one more reason Porcupine helps people say it themselves rather than writing it for them.

What we don’t know yet

Research on messages written with AI help is new and changing fast. We’re watching how people feel about receiving them.

See it in a conversation

The research

From practice

The first professional standard for mediators is self-determination: the parties reach their own decisions, freely and informed. The mediator doesn’t decide for them.

American Arbitration Association, American Bar Association Section of Dispute Resolution, & Association for Conflict Resolution. (2005). Model Standards of Conduct for Mediators, Standard I: Self-Determination.

Studied with adults in general

In small claims court, defendants were nearly twice as likely to fully carry out an outcome they had agreed to in mediation as one a judge imposed.

McEwen, C. A., & Maiman, R. J. (1984). Mediation in small claims court: Achieving compliance through consent. Law & Society Review, 18(1), 11–49.

Studied with adults in general

In experiments, AI reply suggestions made conversations faster and more positive, but people who believed their partner was using them rated that partner as less cooperative and felt less close to them.

Hohenstein, J., Kizilcec, R. F., DiFranzo, D., Aghajari, Z., Mieczkowski, H., Levy, K., Naaman, M., Hancock, J., & Jung, M. F. (2023). Artificial intelligence in communication impacts language and social relationships. Scientific Reports, 13, 5487.

More research

Studied with co-parents

In a study that randomly assigned separating parents to settle custody in mediation or in court, the parent the children didn’t live with was still more involved in their lives 12 years later if the family had mediated, without more conflict between the parents.

Emery, R. E., Laumann-Billings, L., Waldron, M. C., Sbarra, D. A., & Dillon, P. (2001). Child custody mediation and litigation: Custody, contact, and coparenting 12 years after initial dispute resolution. Journal of Consulting and Clinical Psychology, 69(2), 323–332.

From practice

Transformative mediation, a widely used approach, measures success by whether each person leaves stronger and better able to understand the other, not by whether they reach a deal.

Bush, R. A. B., & Folger, J. P. (2005). The Promise of Mediation: The Transformative Approach to Conflict (Rev. ed.). Jossey-Bass.

Theory

Researchers define AI-mediated communication, where AI changes or writes messages on a person’s behalf, and set out how it can change what people say and how much others trust them.

Hancock, J. T., Naaman, M., & Levy, K. (2020). AI-mediated communication: Definition, research agenda, and ethical considerations. Journal of Computer-Mediated Communication, 25(1), 89–100.

Library last updated October 2026.