Project Diana: Building a Better DMCA & Deepfake Takedown System for Creators Who Are Sick of Ineffective, Inhuman & Expensive Subscription Services
Let’s be honest: the internet is full of people making money off other people’s work, bodies, likenesses, and of course the fear of others. Some steal content outright, while others might impersonate them. Some spin up mirror pages, spoof URLs, and deepfake trash like digital raccoons in trenchcoats.
When creators try to fight back we’re often rewarded with confusing forms, vague platform policies, expensive subscriptions, and customer support that feels spiritually outsourced.
The Project Diana Prototype
I’m developing what is currently known as “Project Diana” to assist in addressing common pain points for content creators:
- Subscription-based services
- Expensive most of the time
- No human support
- Ineffective services
The concept is an optional pay-per-use OR subscription-based takedown platform designed for tech-savvy creators who need real protection without being trapped into yet another annoying monthly subscription they barely use.
Subscription fatigue is already a documented issue in digital media, and creators do not need one more recurring charge nibbling at income that is often a lot less glamorous than people assume.
At the center of this project is a pretty simple belief: protection should scale with actual need.. not with fear or an obsession with profiting off of the vulnerable. And most CERTAINLY not with a pricing model that quietly assumes every creator has enterprise money lying around. Public discussion about the “creator economy” loves to act like everyone with a following is rolling in Benjamins, but the harsh reality is wildly uneven, and the middle market is where a lot of people get ignored.
Harvard has already written about The creator economy’s missing “middle class,” and broader research continues to show that creators face serious financial pressure, burnout, and instability even while the sector itself grows.
...So no, the answer is not “just pay another company hundreds of dollars every month forever and hope it fixes your problems.”
The answer should be better tools, outreach, education, DIY-resources, data accessibility, and a system that doesn’t punish creators for needing help only when things go sideways.
Okay.. and? Why does it matter?
DMCA and content-removal isn’t some cute little admin task.. It’s legal, technical, repetitive, time-sensitive, and often emotionally brutal that can often leave people who have to deal with self-removal feeling hopless and defeated.
The U.S. Copyright Office’s Section 512 Study made it clear that the current notice-and-takedown system places substantial burdens on copyright owners and has become deeply contentious and burdensome in practice. Google’s anti-piracy reporting “option” has likewise shown the scale of the problem, documenting billions of URLs removed from Search and explaining that valid copyright notices can affect how infringing sites are surfaced.
In other words this is not a niche annoyance, it’s industrialized abuse, and dealing with it effectively not only requires cleaner evidence, but cleaner workflows and WAY less guesswork for victims the way it’s typically handled.
Where this project comes in & how I intend to address the issue
I’m developing it through a mix of manual takedown research, data extraction, my own method of scripting, database design, web development, and legal-process study. The goal is to turn a chaotic, emotionally loaded process into something more structured and usable, i.e. a system that can:
- Collect and normalize URLs
- Catch duplicate and spoofed links
- Organize evidence
- Track complaint outcomes
- Provide REAL self help soluations that aren’t motivated by money
- Help creators understand what is actually happening instead of tossing reports into the void and praying to the platform gods things turns around.
Lumen exists in part because legal removals on the internet are important to document and analyze, and even Lumen’s own materials reflect how much complexity sits inside these systems. If the process is messy, the software should work harder so the users and researchers don't have to; and so that those with more powerful voices can allocate resources more effectively.
The urgency has gotten worse with impersonation and deepfakes. The FTC reported that imposter scams were the most frequently reported fraud category in 2025, with over 1 million reports and more than $3.5 billion in losses. That is not creator-specific, but it tells us impersonation is not fringe behavior anymore. It’s profitable as hell.
Add gene AI to and we’ve got a nasty epidemic-level petri dish to deal with. Harvard has written about how explicit AI deepfakes have outpaced older revenge-porn laws, leaving victims stuck in legal gray zones while platforms and lawmakers play catch-up. Research highlighted by the Berkman Klein Center also found major differences in takedown success depending on whether reports were made through copyright channels or through platform-specific non-consensual-intimacy policies. That should litreally bother everyone, considering it can affect anyone. Victims should not need a law degree and a blood sacrifice to figure out which reporting lane might maybe work this week.
So the mission here is bigger than automating complaint forms. It is to make takedowns less intimidating, more accessible, and more affordable, while also being tough enough to deal with repeat offenders, spoof infrastructure, impersonation, stolen content, leaked content, and AI-generated abuse. And yes, I care deeply about creators in general, but I care especially about protecting women and children, who are disproportionately targeted by image-based sexual abuse and identity-based exploitation online. Harvard commentary has pointed to findings that deepfake pornography overwhelmingly targets women and girls, and the Cyber Civil Rights Initiative has long documented the severe emotional, reputational, financial, and safety harms tied to nonconsensual intimate-image abuse. This project is being built with that reality in mind. Not as a sterile tech experiment. As a defense tool. Because frankly, there are already enough systems in the world that are easier for offenders to use than for victims.
And if that sounds angry, then good. People deserve tools that respect their work, time, safety, intelligence and creativity. We all deserve better than vague promises, bad scripting, overpriced subscriptions, and fake sympathy stapled to weak execution and incompetent engineers.
I'M BUILDING THIS BECAUSE DMCA AND ABUSE REPORTING SHOULD NOT FEEL LIKE A PUNISHMENT FOR BEING VICTIMIZED IN THE FIRST PLACE, AND IT IS CLEAR OUR GOVERNMENT EITHER LACKS INTELLIGENCE, CAPABILITY, RESOURCES OR A COMBINATION OF ALL THREE WHICH PREVENTS US FROM FEELING SAFE ONLINE.
About the Author
Rae Soria is a technical SEO and marketing professional with experience in digital strategy, search performance, workflow analysis, and content systems. She is a 5 year veteran in content creation, and over the past year, she has also carried out manual DMCA takedowns and deepfake mitigation to better understand offender behavior, complaint workflows, enforcement limits, platform capabilities, and the patterns that matter when building databases, writing scripts, and developing better software for creators under attack.
She has about 10 years of front-end development experience and is currently studying programming at University of Michigan to assist in this project. You can follow her GitHub, although most content will remain private without any notice of intent to collaborate in order to preserve the integrity of methods, resources and tools to building this platform. (In other words, I’m not trying to hand out my work to a bunch of basement dwellers.)
FAQ
What is the Diana Project, exactly?
It is a pay-per-use takedown software platform being developed to help creators respond to stolen content, leaked content, impersonation, and deepfake abuse without getting locked into another monthly subscription.
Why pay-per-use instead of subscription?
Because a lot of creators do not need the same volume of enforcement every single month, and forcing everyone into a flat recurring fee is a great way to make protection feel like a punishment. The goal is to make support scale with real need, not panic-based pricing. Broader digital-media research shows that subscription fatigue is already a real issue for consumers.
Who is this for?
It is built for tech-savvy creators who want more visibility into the takedown process, more control over their data, and a more affordable option than traditional full-service monthly vendors.
What kinds of abuse is it meant to address?
The focus includes stolen or leaked content, impersonation, repeat offenders, spoof URLs, mirror pages, and AI-generated deepfakes. The system is intended to help creators identify, organize, and act on these issues more efficiently.
Why not just rely on existing reporting tools?
Because existing systems are often fragmented, inconsistent, slow, and hard to navigate. The U.S. Copyright Office has acknowledged serious friction in the current notice-and-takedown framework, and recent Harvard-linked research suggests platform reporting outcomes can vary significantly depending on which mechanism a victim uses.
Is this a legal service?
No. It is a software and workflow project, informed by legal consultation and public copyright procedures, but it is not a replacement for legal advice.
What makes this different from competitors?
The goal is to be more transparent, more technically rigorous, and more accessible. That means stronger attention to data quality, spoof detection, URL normalization, repeat-pattern tracking, and a user experience that does not make people feel like they need a decoder ring just to defend themselves.
Why is there such a strong focus on women and children?
Because image-based abuse, impersonation, and sexualized synthetic media disproportionately harm women and girls, and the consequences are not abstract. They are reputational, financial, emotional, and sometimes physical. Any serious protection tool should treat that as central, not secondary.
References (some may not have been cited in-line or in-text)
[1] Deloitte. 2025 Digital Media Trends. Deloitte Insights, 2025.
[2] Harvard T.H. Chan School of Public Health. “Content creators are struggling with mental health, study finds.” 2025.
[3] U.S. Copyright Office. Section 512 Study. Washington, D.C., 2020.
[4] U.S. Copyright Office. “Section 512 of Title 17: Resources on Online Service Provider Safe Harbors.”
[5] Google. How Google Fights Piracy. Google, 2018.
[6] Google. YouTube Copyright Transparency Report.
[7] Lumen Database. “About / Database Overview.”
[8] Lumen Database. “Over Thirty Thousand DMCA Notices Reveal an Organized Attempt to Abuse Copyright Law.” 2022.
[9] Federal Trade Commission. FTC Testifies before the Joint Economic Committee on the Agency’s Efforts to Combat Fraud. March 25, 2026.
[10] Federal Trade Commission. “New FTC Data Show a Big Jump in Reported Losses to Fraud to $12.5 Billion in 2024.” March 10, 2025.
[11] The Harvard Gazette. “How AI deepfakes have skirted revenge porn laws.” January 28, 2026.
[12] Berkman Klein Center for Internet & Society. “Reporting Non-Consensual Intimate Media: An Audit Study of Deepfakes.” September 18, 2024.
[13] Harvard Kennedy School, Carr-Ryan Center. “We Have a Right Not to Be Generated.” January 27, 2026.
[14] Cyber Civil Rights Initiative. “FAQs.”
[15] Harvard Business Review Education / Harvard Business Publishing. The Creator Economy Needs a Middle Class.
