This About page is a work in progress.
AI Usage Policy
The Red String Wire believes in transparency and this is especially important when covering organizations that thrive on its absence.
What AI Does Here
A one time image generation
Text-to-Speech narration.
That’s it.
Every word in this publication is researched, written, and edited by real human beings. Consistent AI usage is for one purpose, and one purpose only: converting those written words into audio.
Why TTS Instead of a Human Voice
The first investigation covers one of the most powerful surveillance companies on Earth. An organization with deep ties to intelligence agencies, law enforcement databases, and government contracts worth billions.
There are many more investigations planned and the targets of those investigations may not be appreciative of the OSINT techniques used, and then having the subsequent findings of those techniques published.
Therefore the authors’ anonymity is not optional, and TTS is a safety measure.
Using a human voice, even if that is a hired narrator, creates both a paper trail and an identity vector. Voice biometrics, accent profiling, and speaker identification are real capabilities deployed by the very entities this publication investigates. Leveraging TTS eliminates those vectors entirely.
The authors are well are aware of the irony of leveraging generated speech to investigate companies that profit from AI driven surveillance. We however consider this an appropriate, and ironically hilarious countermeasure, not hypocrisy.
We would love to be able to use our own human voices instead of fighting with TTS quirks. Not only would it sound better, but it would take half the time, but that’s not a risk we’re willing to take.
How the TTS Works
The narration is generated entirely on-prem using fully open source voice synthesis software via a single NVIDIA RTX 3090 that was originally used for PC gaming and was repurposed for this project.
This means:
- No cloud AI services are used. No audio is sent to OpenAI, Google, Amazon, or any third-party API. The full pipeline runs on hardware owned and operated by the authors with the only data leaving our site being the finished product hosted on Substack and wherever the Podcast is available.
- No corporate AI is subsidized. Your listens do not generate training data for any AI company, nor for us internally.
- No third-party data collection. The TTS process happens on a local machine running in a home office. No telemetry, no analytics, no usage tracking.
Sustainability
The hardware powering this narration runs entirely on solar and the carbon footprint of generating these episodes is approximately equivalent to charging a laptop.
We mention this not to virtue signal, but because we believe it matters to know the full cost of the things you consume. This is a principle that those we investigate have never, and will never embrace.
What AI Does NOT Do Here
AI does not write the articles
AI does not conduct the research
AI does not generate the OSINT analysis
AI does not choose the editorial direction
AI does not draft, edit, or fact check any content
The research draws from public records, SEC filings, FOIA documents, DNS records, court proceedings, open source intelligence tooling, and original analysis. Every claim we make is sourced and everything is documented.
Our Position
We believe in sustainability and transparency in the usage of AI
AI is a tool, and like any tool what matters is who wields it, how they wield it, and whether they are transparent in its usage. We use AI narrowly, transparently, and sustainably on infrastructure we control.
If you do have any questions about our methodology or AI usage, please reach out.
Independent. Unsponsored. Uncontrolled.
