🤫hussh
Agent One
Products
PuppyTagShop
For Business
For Advisors & RIAFor BrandsFor Agent BuildersPartner with Hussh
Blog
LatestProduct UpdatesResearchFounder’s Notes
Glossary
A-ZConsent & PrivacyAgents & AIData Ownership
Company
Our StoryTeamCareersPressContact
Get Agent One
Agent One
Products
PuppyTagShop
For Business
For Advisors & RIAFor BrandsFor Agent BuildersPartner with Hussh
Blog
LatestProduct UpdatesResearchFounder’s Notes
Glossary
A-ZConsent & PrivacyAgents & AIData Ownership
Company
Our StoryTeamCareersPressContact
🤫 University Relations · Seattle

Fifteen minutes from campus. Building the consent-first future with UW.

What we would like to build with the University of Washington: research on personal agents people can trust, research on investing people can check, and paid work for UW students without leaving home.

Start a conversationThe two tracks
Where things stand

Working together toward a partnership. Nothing is signed yet.

We are talking with people at UW and would like those conversations to become a research partnership. Until an agreement is signed, UW is not a partner, and we will not call it one.

This page is ours, not the University's. It describes what we would like to build and does not imply that the University of Washington endorses 🤫 hussh or any of our products. Where we name a UW school, lab or office, it is to say where the work would live.

Why UW

The right people, a short drive away.

Fifteen minutes away

Our home is in Kirkland. A student can spend a summer at the Garage without moving house, and a professor can visit for the price of a coffee. Proximity compounds, and nothing else we do can substitute for it.

Both disciplines, one campus

Consent-first agents need two kinds of rigour: security, privacy and policy on one side, and disciplined quantitative finance on the other. Very few universities have both at world-class depth. UW does, on the same campus.

The same values

UW's Tech Policy Lab was founded to bring technologists, lawyers and the public interest into one room. Making personal data something a person owns and controls, governed by consent, is exactly that kind of work.

Seattle should write this chapter

This region built the cloud, modern online shopping, and a great deal of modern AI. Giving people real ownership of their data in the age of agents belongs here too.

The research

Two tracks, one partnership.

One umbrella, two groups of faculty, each engaging on their own terms. It is how our signed labs work today, and it avoids forcing overlap where there is none.

Track A

Consent-first personal agents: security, privacy and policy

🤫 Agent One acts for one person, on hardware they own, under their consent. The hard questions are about trust: how consent is asked, proven, and kept.

Where it would live at UW
  • Security and Privacy Research Lab, Paul G. Allen School of Computer Science & Engineering
  • Tech Policy Lab (Allen School, Information School, and School of Law)
  • Information School
Questions we would like to answer together
  • What does a rigorous, machine-verifiable consent protocol for AI agents look like, and can an open one such as PCHP become a standard the way SSH did for trust between machines?
  • How should a personal agent ask a person for consent without fatigue, pressure, or dark patterns creeping in over time? What are the right defaults?
  • When a person can swap out their own models, storage and compute, what are the security properties of the whole system, and where are the attack surfaces?
  • How could anyone measure, and publish, how often a company actually tells people when their data is used?
  • If consent-first personal data becomes the norm, what should policymakers hear from independent researchers about how to govern it?
Track B

Disciplined, transparent investing for households

An agent acting for a family will be asked about money. The research question is whether simple, rules-based investing can be tested openly, explained plainly, and audited by the person it serves.

Where it would live at UW
  • Computational Finance & Risk Management program, Department of Applied Mathematics
  • Foster School of Business, Finance and Business Economics
Questions we would like to answer together
  • Over ten and twenty years, do simple, rules-based strategies on large, cash-generating companies hold up once costs, taxes, and ordinary human behaviour are counted honestly?
  • How do simple rules compare with more sophisticated models when both are measured the same way, net of everything?
  • Which benchmarks are honest for long-term household investing, and when does a strategy deserve a new one?
  • How should an agent explain an investment decision so the person it serves can check it, understand it, and say no?
  • Can we build an open-source teaching implementation that students extend as capstone projects, with the results published whatever they show?
For UW students

Real work, paid, and close to home.

A paid summer, at home

Internships and the Garage Residency in Kirkland, fifteen minutes from campus. Paid properly, on work that ships.

A capstone that is real

One team, one real problem, one quarter, in either track. We bring the data and the mentorship; the credit is yours.

Research that gets published

Work with PhD students on questions from both tracks, published on the normal academic timeline, whatever the results.

Your own agent, free

🤫 Agent One is free to every American, students included. Your data stays yours.

See open rolesResearch roles
How to start

Start small. Earn the rest.

The same ladder every university gets, in order of commitment. We would be glad to start with the first step.

  1. 1

    A conversation

    Coffee at the Garage, a lab visit, a guest lecture, a paper read together. No paperwork.

  2. 2

    A student project

    One team, one real problem, one term. We bring the data, the mentorship, and a letter of recommendation.

  3. 3

    Sponsored research

    A faculty-led project on a question we define together, usually 12 to 18 months, with one or two PhD students.

  4. 4

    A named fellowship

    Funding for one PhD student a year, in an area we both care about.

  5. 5

    A multi-year agreement

    Several faculty, clear IP terms, and an annual review. This is how our signed labs work today.

  6. 6

    An advisory seat

    One or two faculty with a standing voice in where our research and product go next.

We fund the engineering and the students' time. We never ask for free labour.

What we promise

Promises we make in public.

  1. 01

    Faculty publish freely

    We never ask a researcher to delay, soften, or bury a finding. If our work fails a test, we want to know, and so should everyone else.

  2. 02

    IP stays clean

    What each side brings, each side keeps. Anything new is worked out in good faith, through the university's own technology transfer office.

  3. 03

    Students come first

    Every engagement has to be worth it for a student: mentorship, real data, published work, or a path to a job. Paid work is paid properly.

  4. 04

    No data broker behaviour, ever

    Anything that touches a joint project is consent-first, with a receipt for every access. That is the whole point of the company.

  5. 05

    Given, not sold

    🤫 Agent One is free to every American, students included. We never sell anyone's data, their attention, or their contacts.

  6. 06

    UW's process, UW's pace

    Sponsored research, IP and licensing go through the University's own offices, including CoMotion, on the University's terms.

  7. 07

    Conflicts disclosed

    Any advisory role or compensation for faculty follows UW's conflict-of-interest rules, disclosed in advance and in writing.

  8. 08

    Student records stay protected

    We never ask for student records. Anything a student shares with us is theirs to share, and theirs to withdraw.

What happens next

Three small asks.

01

A first conversation

Forty-five minutes with faculty in each track and with the Tech Policy Lab, on campus or at the Garage in Kirkland. No deck. A whiteboard.

02

One or two curious PhD students

In each track, who might want to spend a summer on one of these questions. Paid properly.

03

Guidance on the right front door

Whether a multi-track agreement belongs with CoMotion, the College of Engineering, the Office of Research, or somewhere else. We will follow UW's lead.

The same approach already works with our 5 signed university partners. See the honest list

Start a conversation

If any of this resonates, write to us.

Your department or lab, what you teach, research or study, and which track interests you. Faculty, staff and students are all welcome, and every message is read by a person.

partners@hushh.ai