Introducing our Massachusetts Hearing Transcripts service
Legislata is designed to do three things:
Keep up with what’s happening in the political and legislative world
Help you work more efficiently within your own office and with your audience/stakeholders
Do the above without breaking the bank
To this end, users can track legislation in all fifty states and Congress for free, get alerts on press releases from Members of Congress, and set up their own feeds for others to follow (get in touch at chris@legislata.com for help setting that up).
We’re now adding another source of content that we’re maintaining, since we’ve heard so much about how difficult it is to access them: AI-generated transcripts of committee hearings at the Massachusetts State House. And we’re offering it for only $10 per month and free to anyone with a .edu or .gov email address.
How it works
The process we use is simple. When the State House posts a video of a committee hearing, we transcribe it using AI - usually within an hour - and post it here. That means that we can have transcripts much more quickly and cheaply than a traditional transcription service. You don't need to wait around for a transcript to be posted if you need to file a story that night or prepare for a meeting the next day and you don't need to make much room in the budget to have access to what's being said at on Beacon Hill.
That being said, it does rely on the legislature posting the video. If the hearing recording is not available from the State House, then at the moment we don't have it (if you have the recording, then we can transcribe it for you. Get in touch and we can discuss how we can help).
AI-generated quality
The downside of an AI approach is that some of the words will be incorrect, as the AI grapples with a Boston accent or poor audio quality.
To mitigate that, we color-code the transcript. Words in black have a higher than 90% confidence in accuracy. Words in purple are greater than 50% confidence. Words in red are less than 50%. We include the link to the video at the top of every post so you can jump to the timestamp if there's a crucial section that has a lot of red words.
Our AI service doesn't tag speakers by their names. But we have a plan for crowd-sourced speaker identification we hope to roll out by November.
Alternate AI-generation
If the program doesn't work (which can happen if the video is too long or oddly formatted) we'll aim to post it via manually uploading the video to Otter AI. This has the benefit of automatically tagging some speaker names, but it doesn't color code. Each transcript made by Otter will be flagged as such at the start of the transcript.
Cost
This office has a subscription fee of $10/month. It is free for .gov and .edu email addresses and we offer discounts for group purchases (10% for 10+ users, 20% for 50+ users). When you request to join this office you'll be automatically accepted, and we'll send you an invoice, to be paid within one week. For .gov and .edu accounts, there will be no invoice.
Usage
Please feel free to use the transcripts generated here in any professional setting that would not be considered replicating our work - ie, some passages can be shared but not large chunks of the transcript. If you wish to use all of the transcript as a basis for commercial work, please be in touch about licensing. We are also happy to work with you if you’d like to build a paywalled set of more-edited transcripts and host them on Legislata
One more thing…
If you’ve made it this far, you must really be interested in what happens in committee hearings. As a thanks for reading, here is the average sentiment of all hearings across the State House in 2023. As you can see, hearings tend to start off positive, then have a significant drop around when public comments start being made, and then pick up again before dropping off at the close. I’ll leave it to those who have sat through many hearings to say whether the data matches their own experiences.
