Privacy Policy

Last updated: 22 Sep 2026

This policy explains how GenerativeAI Community handles personal information when you visit genaicommunity.ai, apply for membership, or contact our team. The sections on LinkedIn applications and AI assessment apply when you use those features.

Information you share with us

You can browse public community resources without creating an account. When you apply or use an interactive service, we collect the information needed to provide it:

• LinkedIn sign-in: your LinkedIn account identifier, name, profile photo where available, email address, and whether LinkedIn reports the email as verified. LinkedIn asks you to authorize the requested access.
• Bug reports: optional plain-text reports from your agent, stored with your application for admins to review. Do not include credentials or sensitive personal information.
• Membership applications: your contact email, self-reported public LinkedIn profile URL, current or recent role, AI project or use case, your contribution, and reason for joining. You supply your role and project information. LinkedIn sign-in does not provide those details.
• WhatsApp contact: the phone number you provide, including country code, for community membership administration. You supply this number. We do not check whether it is registered with WhatsApp. It is visible to community administrators and included in your emailed application copy. It is not sent to the AI assessor.
• Application administration: consent records, submission and status timestamps, assessment evidence, decision reasons, administrator actions, and email acceptance or failure records.
• Support and other services: information you send when contacting us. If you submit a job listing, we collect the listing and contact details you provide; published listing fields are visible to visitors. Where payment is offered, the payment provider processes payment details and we receive transaction records.

Please avoid including confidential employer information, unnecessary sensitive information, or personal information about other people in an application.

Technical information and cookies

Our hosting and security providers process connection information, such as IP addresses and request metadata, to deliver the website and protect it from abuse. The application uses a hashed IP address for login rate limits. Necessary cookies support sign-in, sessions, and security checks. Blocking these cookies may prevent sign-in.

The membership application does not need advertising cookies. Following a link to another platform may expose you to that platform’s cookies and tracking practices.

How we use information

• Authenticate applicants using LinkedIn-provided identity and email, and prevent duplicate applications and abuse.
• Assess membership applications and support administrator review.
• Send submitted application copies, application updates, and approved membership invitations.
• Operate community services, respond to requests, and investigate technical or security problems.
• Meet applicable legal obligations and handle disputes.

We request consent before an application is submitted or agent access is issued. Membership consent is not consent to unrelated marketing. Our legal basis for processing depends on the service and applicable law. It may be your consent, the need to provide a service you request, or a legal obligation.

AI-assisted membership assessment

When enabled, our AI assessor on Cloudflare reviews your application for evidence that you build, research, or use AI in practice. Exceptional student projects can qualify. Clear matches can be approved automatically; unclear applications and failed assessments go to a community administrator. Students without exceptional evidence can be declined automatically; approved company affiliations remain an exception.

We retain relevant evidence and reasons with the application decision. An administrator can approve or decline an application with a reason. Rejected applicants whose original submission used an agent can appeal with work evidence, a project explanation, or a community reference. Appeals receive human review. We record the evidence, original rejection, appeal decisions, and administrator actions. Named references are not treated as verified endorsements. You can contact us to correct inaccurate personal information. LinkedIn sign-in does not independently verify the background you provide.

Applying through Codex or another agent

After signing in and giving consent in your browser, you can choose to generate a token for an HTTP-capable agent. Anyone with your token can read your application, submit it once, and check its status until the token expires or is revoked. New tokens expire after 30 days; previously issued tokens keep their original expiry. Generating a replacement invalidates the previous token. We store the token’s hash, not the token itself.

We do not share your LinkedIn password or OAuth tokens with your agent. The agent provider’s policies also apply to information you give it. Revoking our application token does not delete copies already held by that provider.

Who receives information

Authorized community administrators access applications to operate admissions and respond to requests. Cloudflare provides hosting, application storage, AI assessment, and transactional email services. LinkedIn provides sign-in. After you save application consent, we may send your verified business email or submitted LinkedIn profile URL to Crustdata to retrieve professional profile details. We show those details to you and may use them as background context in assessment; we do not treat them as proof of your personal contribution. For applicants claiming an ICML, NeurIPS, or ACL main-track paper, we send the applicant name, paper title, venue, and year to Parallel Search to find a public proceedings record. Each service processes the information it needs to perform that task.

If approved, we email a WhatsApp invitation from [email protected]. Joining is your choice. WhatsApp and other members may see your phone number, profile information, and messages according to the group type and your settings. WhatsApp’s own terms and privacy policy apply. Email acceptance does not mean that you have joined the community.

Our existing community rules warn that group messages may be made public or used to train a bot, and that other members may download chat history. Do not treat group conversations as confidential. This community-message notice is separate from the application assessment described above.

External registration, event, job-application, or payment links lead to services with their own privacy policies. We do not sell membership application information or provide it to third parties for their own marketing. We may disclose information where required by law or necessary to address fraud, security incidents, or legal claims.

Member chat archive

We use Cloudflare Workers AI to help route new posts among our community groups. The model may receive the original message text, including contact information written in it, during the live sync. We do not retain the original classification text in the archive, review database, or sync exports. Admins review selected examples with phone numbers removed to improve routing rules and measure accuracy. After a reviewed trial, our linked WhatsApp account may reply to a misplaced post with a suggested group and the signature “Sent by my agent”. Reviewing examples does not automatically train the hosted model.

Past Chats stores selected community group messages in Cloudflare D1 so approved members can search them. The archive includes message text, author display names, group names, and timestamps. Attachments and phone-number metadata are omitted. Phone numbers found in message text or display names are replaced before storage with a stable color, animal, and time-of-day alias; messages may still contain other personal information written into their text. The synced history may be incomplete.

Approved members can access the same archive through a valid application token provided to their chosen agent. Administrators can hide individual messages or unpublish groups. Contact us below to request correction or removal. Removal prevents future access through our archive but cannot erase copies already obtained by members or their agent providers. This archive is separate from the AI membership assessment; message retrieval does not itself train a model.

Retention and security

We keep application and decision records for as long as needed to administer your application or membership, resolve support requests, and meet legal or security obligations. We do not have a fixed automatic deletion period for all application records. You can request deletion using the contact below; limited records may need to remain for a valid legal or security reason.

Sessions, application tokens, and login-state records expire and are cleaned up separately. We use HTTPS, access restrictions, limited-scope tokens, and hashed session and application-token storage. No online service can guarantee complete security. Our providers may process data outside your country; their service locations and applicable legal requirements govern that processing.

Your choices and requests

Contact us to request access, correction, deletion, a usable copy of your information, withdrawal of consent, or information about how an application decision was made. Depending on applicable law, you may also have rights to object to or restrict processing. We may need to verify your identity before acting. Removing required application information can prevent us from processing the application. Withdrawal does not undo processing already carried out lawfully.

Contact us if you believe a child has submitted information without appropriate authorization so we can review and address it. Please do not submit information on a child’s behalf through the membership application.

Updates and contact

We will update this page when our practices change and revise the date above. Material changes to application consent will be presented in the relevant application flow where appropriate.

For privacy questions and requests, contact Nirant Kasliwal, community administrator, at [email protected]. Our noreply address only sends email. Use this contact address for replies and privacy requests.