Hana Ellery Profile Speaking enquiries

HanaEllery

Founder of Tern Labs. Author of Slow Intelligence.

She builds the tests an AI system has to pass before it is allowed to meet the public.

Hana Ellery drawn in graphite points: grey hair, round glasses, a striped top, arms crossed

Twenty years of asking machines to show their work.

Hana trained as a computational neuroscientist in Edinburgh and spent a decade leading speech research before co-founding Tern Labs in London in 2014. She lives between London and Singapore, and still reviews every new test suite before it ships.

2014
Tern Labs founded in London
61
Regulated organisations on Tern
140
Keynotes in 31 countries
The point portrait turned to three quarters, showing it is a volume

Talks that leave a room asking better questions.

Since 2016 Hana has given 140 keynotes in 31 countries, from engineering conferences to central bank retreats. Her talk Machines That Ask First has been watched 4.2 million times.

  • Machines That Ask First Keynote, 45 min
  • What an audit cannot see Talk, 30 min
  • Evaluation is a product Workshop, half day
A speaker on a bright stage above the silhouettes of an audience, drawn in points

Tern tests AI systems the way engineers test bridges.

Tern Assay runs thousands of scenario tests against a model before release and writes up the evidence in a form a regulator can read. Hospitals, banks and public agencies use it to decide what goes live.

2.3 bn
Test cases run in 2025
180
People in London and Singapore
Series C
Raised in 2024
The Tern Assay release check screen drawn in points: 97.8 percent of tests passed

Slow Intelligence, on knowing when a machine should wait.

Published in 2023 and translated into 14 languages, the book argues that the most useful AI is the one able to say it is not sure, and it shows the institutions that make that possible.

Read the first chapter 380,000 copies in print
An open hardback book drawn in points, its pages fanning from the spine

Advice given in 23 countries, from ministries to boardrooms.

Hana has sat on evaluation panels for nine national regulators and chairs the independent Model Evidence Group, which publishes open test standards for public sector AI.

  • London
  • Singapore
  • Brussels
  • Stockholm
  • Nairobi
  • Cape Town
  • Abu Dhabi
  • Bengaluru
  • Seoul
  • Tokyo
  • Sydney
  • São Paulo
  • Toronto
  • San Francisco
A dotted world map with arcs from London to fourteen cities

A career built one tested claim at a time.

Six moments, each a claim she had to prove before anyone else relied on it: a thesis, a lab, a company, a hospital, a book and a standard.

22
Years in machine learning
14
Languages the book is in
9
Regulators advised
  1. 2003PhD in computational neuroscience, Edinburgh
  2. 2009Head of speech research, Halden Institute
  3. 2014Co-founds Tern Labs in London
  4. 2019First Tern Assay release, hospital triage
  5. 2023Publishes Slow Intelligence
  6. 2025Chairs the Model Evidence Group
A timeline drawn in points with six years marked from 2003 to 2025

Bring Hana to your stage, your board or your inbox.

Speaking enquiries

Keynotes, panels and board sessions, booked through Clara Whitlow at Northlight Speakers, London.

Field Notes, the newsletter

One letter a month on testing AI in the real world, read by 62,000 people.

Press and research

press@hanaellery.com
Tern Labs, Clerkenwell, London EC1
The point portrait of Hana Ellery again, now in full colour