I would be interested to read and review the minimum standards that you'd like to advocate for. It would be great if this was also presented alongside best practices so orgs can plainly see what to aim for (which should be above minimum standards).
Everyone should be free from harassment and discrimination, and I think it's fair for us to hold communities and orgs that are aiming to do good to a high standard of behaviour.
❤️
Thank you three so much for coordinating this, and more generally for your leadership in making EA a safer place.
We (Frances, Sarah, and Geetha) work in AI safety and the broader EA ecosystem. We are increasingly concerned about sexual harassment, workplace harassment, and discrimination in these spaces. Through our advocacy, we have heard numerous firsthand accounts of ineffective or nonexistent reporting systems and anti-harassment policies, dismissive organisational attitudes, and harassment incidents being mishandled. We are now collecting incident reports through confidential calls, so we can make the case to funders and organisations for minimum standards on how harassment is prevented and handled. If you have an experience to share, we’d love to hear from you (please fill out this form).[1]
Please share this widely: within your organisation, in relevant Slack and fellowship groups, on social media, directly with people you know, and anywhere else you’re comfortable.
We will produce a list of recommendations for how organisations in the space can improve, including policies, guidance on handling conflicts of interest in complaints, resources to support victims through reporting, and standards for reporting and codes of conduct at events and co-working spaces.
To make these recommendations effective, we need to understand the issues victims are facing and be able to explain the scope of the problem. We believe a collection of concrete incidents is an important first step.
We plan to advocate for better baseline standards across organisations in the ecosystem by approaching funders such as Coefficient Giving, Longview Philanthropy, FLF, and SFF, and urging them to require minimum standards from their grantees. (Note: we do not have full details on each funder’s existing standards, but have been disappointed by what we do know.) We are also planning to draft an open letter to organisations in the space that professionals will have the chance to sign.
We are doing this independently. We are not funded by anyone, and this project is not affiliated with any organisation, including our employers.
We’re asking people to share incidents of sexual harassment, workplace harassment, and discrimination with us. Right now, we’re particularly focussed on AI safety, but we’re also keen to hear about incidents within effective altruism more broadly. These can be firsthand accounts (something you experienced directly) or corroborated secondhand accounts (something you directly witnessed, or an incident you have evidence of). If you’re not sure whether something “counts,” please speak with us anyway. We’d love to hear from as many people as possible.
We’ll ask you to have a call with one of us to record the incident. Nothing you share will go further without your permission. We’ll work with you to decide what, if anything, we can share and how widely (e.g. you might be comfortable with it being shared within the ecosystem but not in a public post, or with naming an organisation but not yourself). You can change your mind at any time. The only exception is if you disclose an imminent threat to someone’s safety (e.g. “I have heard that [Name 1] plans to do [violent thing] to [Name 2]”), in which case we would notify [Name 2].
You’ll need to identify yourself to the person you speak with, so we can confirm that every report comes from a verified source. You can ask for your details to stay anonymous to the other two of us.
(General note: If you’ve previously disclosed an incident to any of us, it will not be included in this project in any capacity by default. If you would like it included, please reach out again and let one of us know.)
(Second general note: At least one other person will help us analyse and write up our findings. Anyone helping at that stage will only have access to information we’ve been given explicit permission to share.)
Any of us! Or all of us. It’s totally up to you. We know you may not feel comfortable talking to any of the three of us, so capacity permitting, we’ll try to offer one or two other people you could speak with later on.
Please fill out this form with, at minimum, your name and email. If applicable, you can specify who you’d like to talk to. Once you submit, we’ll reach out to book a 30-minute call. The only people with access to form submissions are Sarah, Geetha, and Frances. To quickly reiterate, we are working on this project in a personal capacity and not in affiliation with our employers. All views expressed are our own.
Sarah Gokhale
Hello! I am the Events Lead at Coefficient Giving. Previously, I was a Recruiter also at Coefficient Giving and before that I was on the AI Events and EAG teams at the Centre for Effective Altruism. I left CEA because of how badly the org and its leaders mishandled Frances’s sexual harassment investigation. I gave a critical exit interview about this that was never escalated to leadership.
I got into effective altruism while doing a PhD in moral philosophy at the CUNY Graduate Center in New York City.
I moved to the Bay Area in August 2026 from Portland, Oregon. Since then, I have been motivated to speak increasingly publicly about the sexual harassment and discrimination I witness in the AI safety field on a weekly basis. I am eager to speak candidly with and be a resource to anyone who would find it useful.
Geetha
Hello :) I’m a PhD student at UC Berkeley working on biosecurity, specifically transmission suppression / indoor air quality interventions. I’ve been vegan since I was 14, and I learned about effective altruism in my second year of undergrad. After that, I began exploring global health and animal welfare work, and later transitioned into biosecurity. Before my PhD, I completed my Master’s in the Sculpting Evolution Group at MIT under Dr. Kevin Esvelt, and I’ve also worked at SecureBio.
Over my 8 years in the community, I’ve grown increasingly concerned about incidents of harassment and discrimination, and about attitudes towards them. I supported Frances throughout reporting sexual harassment to CEA, including attending calls alongside her. I was shocked by how much self-advocacy it took to get any help and by how visibly uncomfortable and inexperienced senior leaders were throughout the process. The stories I have heard from members in the community since this incident have been equally as concerning.
Frances
Hi! I'm Frances, and I'm a Project Manager for Events at FAR.AI. Before that, I worked on the EA Global team at the Centre for Effective Altruism, and before that, I was an Operations Associate at AI Safety Support.
I've previously written about the sexual harassment I experienced at the Centre for Effective Altruism, which was mishandled and ultimately ended in a settlement. While I am by no means an expert, I have firsthand experience navigating reporting systems and understand how terrible it can be.
Through this project, I want to give victims a chance to share their experiences without having to put themselves in a stressful or unsafe position. More than anything, I care about victims and want to support however I can. I also believe AI risk is an incredibly important problem, and the field desperately needs talent. For people to do their best work, they need to be able to work safely. I want our professional ecosystem to be as healthy and functional as possible.
I posted a different form link on X. I am still checking that one, but we're planning to predominantly use this new form link moving forward. I am the only one with access to previous form responses, and as stated, those will not be shared with anyone else.
Thank you for doing this. It takes a lot of courage, especially when you’ve been through it yourselves.
I’m not sure if this is useful, but sharing in case it helps: the international development and humanitarian sector went through something similar after the Oxfam sexual exploitation scandal in Haiti came to light in 2018. Donors like FCDO started requiring safeguarding policies, reporting mechanisms and staff training as part of due diligence before funding larger organizations and their local partners.
Here's the link:https://www.gov.uk/guidance/safeguarding-and-protection-from-sexual-exploitation-and-abuse-and-harassment-seah-in-the-aid-sector
When I managed grants and sub-grants at the International Rescue Committee, we had to meet those standards and make sure that our local partners did as well. The system isn’t perfect, but it could be a useful starting point for what funders can realistically require from grantees.