About us

T3C is built by the AI Objectives Institute, a nonprofit R&D lab dedicated to decoding and realigning the incentives that determine AI futures. Our technology is open source and you can run it yourself by visiting our Github page. We are always open to collaboration and you are welcome to reach out to us at t3c@objective.is.

Our mission in building T3C is to make it as easy as possible for groups to understand one another. We are entering an era in which it is cheaper to manufacture data about people than to listen to their voices. At AOI, we believe that building technology to capture and represent human perspectives is a critical step towards preserving our agency and identities as constituents, community members, professionals, students, patients, decision-makers...and humans.

General FAQ

Do you collaborate?

Yes! We're always open to new projects and funding opportunities. We have a wide range of research interests and goals for T3C, so if there's something we don't currently do that you need, please reach out to discuss.

What's your track record?

We've generated hundreds of reports and collected data from thousands of respondents for projects with governments, institutions and communities, including:

Taiwan: We partnered with the Ministry of Digital Affairs to analyze public input on AI policy, same-sex marriage, and 2024 election platforms. Taiwan now hosts its own T3C instance, expanding to national institutions.
Tokyo Elections: We engaged 1,000 residents to identify policy priorities, informing the 2024 gubernatorial race.
6th European Rural Parliament in Scotland: We partnered with CrownShy and Scottish Rural Action (SRA) to support the 6th European Rural Parliament (ERP) in Inverurie, Scotland in 2025.
Yemen: CMI - Martti Ahtisaari Peace Foundation deployed T3C in 2025 to safely gather youth views on political participation and peacebuilding across 18 governorates with a 94% completion rate.

Your use cases resonate but AI scares me. Should I still get in touch?

Yep. We think a lot about how, when and why to apply AI to data collection and analysis. We'll always be honest about what our tool can and can't do, and we want to know what you think!

Product FAQ

Where can I get my privacy and security questions answered?

We have a detailed privacy and security policy. If you still have questions or specific requirements we can't meet, please reach out at t3c@objective.is. We are actively working on supporting multiple LLMs and welcome any technical or financial support towards that goal.

Do you plan on supporting elicitation through platforms other than WhatsApp?

Yes, we are working on adding browser chat currently. If you have other platforms you're interested in, please reach out.

What phone numbers do you use to send WhatsApp messages?

Our default is a U.S. phone number. You can provide your own non-U.S. phone number when setting up the study. If you would like us to acquire one for you, please reach out. Also, if you find that your number isn't working, please let us know so we can work with you on it.

Do I have to collect data with your tool in order to generate reports?

No, you can bring your own data for reports. T3C works with the following column headers:

“comment” contains the text of the responses to analyze with T3C (required)
“interview” contains the respondent name, which can be a pseudonym, or blank for anonymous speakers (optional)
“id” for linking to any external data, will be generated if blank (optional)
“video” and “timestamp” for links to Vimeo videos and the timestamps of specific text (optional)

A sample CSV is available to download.

How can I make an audio transcription?

T3C can process transcripts from audio or video recordings. You can use cloud-based services like Otter.ai, Rev, or Happy Scribe, or desktop applications like Descript or OpenAI Whisper for local transcription. We have a detailed transcription guide covering these options and their trade-offs.

Are there any language limitations?

T3C can process text responses in 50+ languages, and audio messages are transcribed using OpenAI Whisper. Whisper generally performs well across major languages, but accuracy can vary for certain accents, regional dialects, or low-resource languages, as well as in noisy recording environments. This means some transcriptions may contain errors. For projects where language accuracy is critical, we recommend reviewing a small sample of transcripts to ensure the model meets your needs.

How do I customize the analysis prompts?

Expand the “Advanced Settings” section in the report creator to find “Customize AI prompts.” You can modify prompts to: