Stevie Bergman

Bio

Stevie is a public interest scientist and technologist, with a global, human rights and ethics focus. She researches questions at the interface of technology and society, rights, and policy. Currently, she is a senior research scientist on the Ethics Research Team at Google DeepMind.

Her primary areas of investigation include: (1) Sociotechnical research for effective technology policy and governance; (2) Data, annotation, and the hidden, often exploited labor underlying AI; (3) The impacts and implications of AI and other emerging technologies for at-risk and marginalized communities in the Majority World.

Stevie has expertise in in human rights and international humanitarian aid, algorithmic fairness, participatory AI design, misinformation/disinformation and technology, multilingual AI, recommender systems, trust and safety classifiers, and social media, science & research practices.

She has over a decade of experience in science & technology communication.

See her CV for more information.

Professional experience

Senior Research Scientist, Ethics Research Team
Google DeepMind
4/2022 - Present

Research Scientist, AI Fairness, Responsible AI
Meta
1/2020 - 4/2022

Lecturer, Ethics of AI Graduate Seminar
Princeton University
Spring & Fall 2021

Fulbright Fellow, Indonesia
Institut Teknologi Bandung
10/2013 – 7/2014

US Peace Corps, Science Education Volunteer
Uganda
2/2011 – 4/2013
Mathematics, physics, and computer teacher at Lacor Secondary School and Warr Girls. Creator & Co-director of GirlTech Uganda and Secondary Education PCV trainer.

Vaccine Litigation Paralegal
US Department of Justice
8/2009 – 12/2010

Other projects include being the creator and co-host of the These Vibes Are Too Cosmic radio show on WPRB Princeton 103.3 FM, creator and producer of a podcast miniseries on AI & Human Rights, weekly tutor at Garden State Prison in New Jersey, and volunteer with the Flatbush Mutual Aid in 2020.

Education

Stevie has a doctorate in experimental physics from Princeton University (2019) where she held an NSF Graduate Research Fellowship and was awarded the Joseph Henry Merit Prize. Her thesis work included measuring the statistical polarization in the cosmic microwave background radiation from the early Universe among instrumental, environmental, and astrophysical noise, via constructing the most sensitive radio telescopes ever built (the SPIDER instrument) and flying it on a balloon above Antarctica.

She has a BA in physics from Smith College (2009), where she graduated cum laude, with Phi Beta Kappa and Sigma Xi. She was awarded the Waterman Prize for Outstanding Senior in Physics.

Papers in Tech & Society

Full list of journal articles available on Google Scholar, including publications in physics.

2024

STAR: SocioTechnical Approach to Red Teaming Language Models
Laura Weidinger, John Mellor, Bernat Guillén Pegueroles, Nahema Marchal, Ravin Kumar, Kristian Lum, Canfer Akbulut, Mark Diaz, Stevie Bergman, Mikel Rodriguez, Verena Rieser, William Isaac

STELA: A Community Centred Approach to Norm Elicitation for Agent Alignment
Stevie Bergman, Nahema Marchal, John Mellor, Shakir Mohamed, Iason Gabriel, William Isaac
Published in Nature Scientific Reports 2024

The Ethics of Advanced AI Assistants: Access and Opportunity (Chapter 15)
A. Stevie Bergman, Renee Shelby, Iason Gabriel (Blog post)

The Illusion of Artificial Inclusion
William Agnew, A. Stevie Bergman, Jennifer Chien, Mark Diaz, Seliem El-Sayed, Jaylen Pittman, Shakir Mohamed, Kevin R. McKee
Pulbished in CHI 2024

2023

Sociotechnical Safety Evaluation of Generative AI Systems
Laura Weidinger, Maribeth Rauh, Nahema Marchal, Arianna Manzini, Lisa Anne Hendricks, Juan Mateos-Garcia, Stevie Bergman, Jackie Kay, Conor Griffin, Ben Bariach, Iason Gabriel, Verena Rieser and William Isaac

Representation in AI Evaluation
A. Stevie Bergman, Lisa Anne Hendricks, Maribeth Rauh, Boxi Wu, William Agnew, Markus Kunesch, Isabella Duan, Iason Gabriel, William Isaac
Published in FAccT 2023

2022

Towards Responsible Natural Language Annotation
for the Varieties of Arabic
A. Stevie Bergman and Mona Diab
Published in ACL Findings 2022

Adaptive Sampling Strategies to Construct Equitable Training Datasets
William Cai, Ro Encarnacion, Bobbie Chern, Sam Corbett-Davies, Miranda Bogen, Stevie Bergman, Sharad Goel
Published in ACM FAccT 2022

Guiding the Release of Safer E2E Conversational AI through Value Sensitive Design
A. Stevie Bergman, Gavin Abercrombie, Shannon Spruit, Dirk Hovy,
Emily Dinan, Y-Lan Boureau, Verena Rieser
SIGDIAL 2022

SafetyKit: First Aid for Measuring Safety in Open-domain Conversational Systems
Emily Dinan, Gavin Abercrombie, A. Stevie Bergman, Shannon Spruit,
Dirk Hovy, Y-Lan Boureau, Verena Rieser
Published in ACL 2022

Fairness On The Ground: Applying Algorithmic Fairness Approaches to Production Systems
Chloe Bakalar, Renata Barreto, Stevie Bergman, Miranda Bogen, Bobbie Chern, Sam Corbett-Davies, Melissa Hall, Isabel Kloumann, Michelle Lam, Joaquin Quinonero Candela, Manish Raghavan, Joshua Simons, Jonathan Tannen, Edmund Tong, Kate Vredenburgh, Jiejing Zhao

TVR2C Radio

Stevie was the creator and co-host (with Frances Kraus) of These Vibes Are Too Cosmic from 2015-2019. Also known as TVR2C, shows aired weekly on WPRB Princeton 103.3 FM, and featured expert interviews from across science and tech. All full streams of past shows are available in the backcatalogue.

Sampling of some favorites:

Ruha Benjamin on Power and Manipulation at the Interface of Tech and Society

Edward Felten on Policy and Tech and Electronic Voting

Annette Zimmerman on Algorithmic Injustice

Bendert Zevenbergen on the Power Dynamics and Ethical Frameworks in AI and the Internet

Joseph Amon on Health and Human Rights

Sam Wang on Gerrymandering and Stats (Show 1 and Show 2)

Alexander Ploss on Research in Infectious Diseases and It's Global Importance

Sébastien Phillipe on Nuclear Arms Verification and Disarmament

AI & Human Rights Podcast

Five episode podcast miniseries primer on AI & Human Rights created in 2019, framed by the joint Princeton University Center for Information and Technology Policy and UN conference Social Protection by Artificial Intelligence: Decoding Human Rights in a Digital Age. Listen on Apple, Spotify, or Soundcloud.

Contact

steviebergman {at} deepmind {dot} com

Engagements

Stevie is an experienced communicator and educator. She regularly speaks at events on the various topics under her expertise. Her teaching experience is discussed in her bio and CV, not to mention her radio appearances and podcast. See below for a sampling of her recent speaking engagements, and fill the contact form if you would like to contact her about speaking at an event.

2024

Keynote speaker, LREC-COLING 2024 Conference Third Workshop on Safety for E2E Conversational AI
Safety for Whom?

Invited speaker, Center for Democracy and Technology
AI Governance, Evaluations, and Representation

2023

Fireside chat, NIST/CASMI workshop
Risk Management (or Measurement), for whom, why and how?

Panel speaker, Cambridge University Disinformation Summit
Potential Benefits and Risks of Generative AI
Full recording on YouTube

Speaker, ACM FAccT Conference
Representation in AI Evaluations

Workshop presenter, Cambridge Many Worlds of AI
Envisioning equitable representation in AI

2022

Panel Speaker, Building the NIST AI Risk Management Framework (Workshop #3)
Recording can be found here.

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';