A documented investigation into automated content plagiarism and what it reveals about AI-powered publishing systems
Disclaimer: This article presents my documented personal experience and
publicly available data analysis. I make no claims about content I have
not personally verified. Q2BSTUDIO is invited to provide their perspective.
The Discovery
On November 30, 2025, I published an article on dev.to announcing a creative challenge for OBSIDIAN Neural, my open-source VST3 plugin for AI music generation. The project had been presented at AES AIMLA 2025 in London in September, and I wanted to engage the community with this competition.
A few days later, while checking new user registrations on my platform, I noticed several users with @qq.com email domains. Curious about potential new interest, I decided to search Google for mentions of OBSIDIAN Neural.
After scrolling through several pages of results (around page 7 or 8), I found something unexpected: my article had been translated into Spanish and published and on q2bstudio.com, a Barcelona-based software development company that sells AI services to businesses.
What I Found
The Reproduced Article:
- Approximately 80% of my article was translated and reproduced
- Published on the same day as my original (November 30, 2025)
- No attribution to me as the original author
- No link back to my dev.to post
- Published under their blog section promoting their AI services
The Modifications:
This wasn't just a simple copy. The content had been actively modified:
- Hyperlinks were injected into the text redirecting to Q2BSTUDIO's commercial services:
- "IA para empresas y agencia de IA" (AI for businesses) linked to their AI services
- "aplicaciones a medida y software a medida" (custom software) linked to their development services
- A paragraph titled "Cómo participar y prerequisitos" (How to participate and prerequisites) was added, mentioning Q2BSTUDIO and their services
- A paragraph "Sobre Q2BSTUDIO" (About Q2BSTUDIO) was added at the end, presenting their company
- Their marketing pitch was integrated throughout
The Deceptive Presentation:
The article's wording suggested Q2BSTUDIO was involved in organizing my independent challenge. Phrases like "como sigue siendo una iniciativa en planificación nos gustaría saber si te interesaría participar" (as this remains an initiative in planning, we would like to know if you would be interested in participating) created confusion about who was actually behind the project.
Important Note:
This was not a random AI "hallucination." The reproduced article correctly linked to obsidian-neural.com, accurately described the challenge prizes, and maintained the core details of my announcement—while inserting their commercial links and branding.
The Scale of the Operation
When I visited their blog on December 6, 2025, I made a striking observation: I manually counted approximately 180+ articles published that day alone across the first 20 pages of their blog.
Their pagination showed over 36,516 pages of blog content.
This volume strongly suggested an automated publishing system operating at industrial scale.
My Response
On Saturday, December 6, 2025, I sent a formal email to Q2BSTUDIO requesting:
- Immediate removal of the article
- An explanation of how this happened
- Clarification about their relationship (or lack thereof) with my project
Their Reply
Q2BSTUDIO responded within hours. Here is their complete response, quoted verbatim:
"Dear Anthony,
We regret what happened. Occasionally, our AI publishes uncontrolled posts that reflect its own bias. The post has been removed.
Saludos,
Charlie"
The article was indeed removed from their website.
The Investigation
Their response, while resolving my immediate issue, raised serious questions. For a company that sells AI services to clients, the admission that their "AI publishes uncontrolled posts" was particularly concerning.
I decided to investigate further. To ensure objectivity and transparency, I developed open-source tools that anyone can verify and reproduce. I developed a Python script (available at https://github.com/innermost47/q2bs) to:
- Systematically scrape their blog
- Analyze publication volumes
- Automatically archive content to Wayback Machine
What the Data Revealed:
Note: My computer crashed during the scraping process, so these numbers represent partial data. However, the findings are already significant:
- Total articles documented: 144,966
- Analysis period: November 20 to December 7, 2025
- Average daily output: 8,401 articles per day
- Peak day: 10,251 articles (December 4, 2025)
- Publication frequency: 1 article every 10.3 seconds on average
I'm currently running a new scraping campaign to gather more complete data.
Comparison with Major Publishers:
For context, major news organizations publish approximately:
- TechCrunch: ~40 articles per day
- The Verge: ~30 articles per day
- The New York Times: ~250 articles per day
Q2BSTUDIO's system is producing more than 30 times the volume of The New York Times.
The Irony
There's a profound irony here: an AI development company using AI to plagiarize content about... AI development. OBSIDIAN Neural is itself an AI-powered tool for music generation. Their automated system copied my announcement about AI music generation to promote their own AI services—while claiming their AI has "uncontrolled bias."
The Company Behind the System
This raises an important question: Who is Q2BSTUDIO?
Q2BSTUDIO is a legitimate software development company:
- Founded: 2008
- Location: Barcelona, Spain
- Services: Custom software development, cloud services (AWS/Azure), cybersecurity and pentesting, business intelligence (Power BI), process automation, artificial intelligence
- CTO: Carlos Pérez Martín
- Recognition: Carlos Pérez participated in La Vanguardia's "Premi Jove Empresari" (Young Entrepreneur Award) campaign in 2017
This is not a fly-by-night operation. This is an established company with a public reputation, recognized by Spanish media, that markets AI solutions to business clients.
The Questions This Raises
For Companies Selling AI Services:
- If Q2BSTUDIO cannot control their own AI systems, how can clients trust the AI solutions they provide?
- What quality control measures should exist for automated publishing systems?
- How can companies prevent their AI from monetizing others' content without permission?
For Content Creators:
- How can independent developers protect their work from automated plagiarism at this scale?
- Should we be periodically checking if our content appears elsewhere without attribution?
- What recourse exists when AI systems plagiarize at industrial volume?
For the Industry:
- As AI content generation becomes ubiquitous, how do we maintain attribution and copyright?
- What standards should exist for AI-powered publishing systems?
- How do we balance automation with accountability?
The Broader Implications
I can only verify my own case. I have not investigated the origin of their other 144,966+ documented articles, and I make no claims about content I haven't personally examined.
However, this experience raises important questions:
- If their AI "plagiarized" my technical article, could it be doing the same to others?
- How many other creators might be affected without knowing?
- What percentage of their massive content volume is original versus reproduced from other sources?
For Other Potentially Affected Authors
If you've written technical content in English, it's worth checking:
site:q2bstudio.com "[your article title or unique phrase]"
If you discover your content on q2bstudio.com without attribution, here's what I learned:
- Document first: Screenshot everything before contacting them
- Be specific: Reference exact URLs and dates
- State your requirements clearly: Removal, attribution, explanation
- Give a reasonable deadline: 48-72 hours is fair
- Keep records: Save all correspondence
Based on my experience, they do respond to direct contact and remove content when requested.
Important: I have analyzed only publicly available information from
Q2BSTUDIO's blog. I make no claims about the origin of articles I have
not personally examined. The questions raised here are intended to prompt
investigation, not to assert conclusions.
Call to Action: For Journalists and Investigators
This case deserves deeper investigation. Questions that warrant professional journalism:
Content Analysis: What percentage of Q2BSTUDIO's 36,516+ blog pages contains plagiarized or substantially reproduced content from other sources?
Pattern Recognition: Are there systematic patterns in how their AI system selects, modifies, and republishes content?
Commercial Impact: How does this automated content generation and link injection affect their SEO rankings and business development?
Industry Scope: Are other companies running similar AI-powered content farms at this scale?
Legal Framework: What are the copyright implications of AI systems that automatically reproduce, modify, and monetize others' content?
I've made my investigation tools publicly available on GitHub. The data is there. The pattern is documented. This story is bigger than one plagiarized article.
The Resolution
Q2BSTUDIO removed my article promptly, and I appreciate their responsiveness to my individual complaint.
However, the fundamental issue remains unresolved: an established company with a public reputation, that markets AI services to clients, operates an automated publishing system that:
- Produces 8,000+ articles per day
- Operates with admittedly "uncontrolled" AI
- Monetizes content through injected commercial links
- Creates confusion about content authorship
The challenge for all of us in the AI space is this: How do we harness the power of automation while maintaining the standards of attribution, quality, and accountability that both creators and consumers depend on?
Technical Documentation
My Original Article:
- Platform: dev.to
- Published: November 30, 2025
- Topic: Creative challenge for OBSIDIAN Neural (open-source AI music VST)
- URL: https://dev.to/innermost_47/planning-a-creative-challenge-for-obsidian-neural-early-interest-check-5c6h
The Reproduced Version:
- Platform: q2bstudio.com
- Published: November 30, 2025 (same day, translated)
- Status: Removed after complaint on December 6, 2025
- Evidence: Screenshots and email correspondence documented
Q2BSTUDIO's response email from "Charlie" dated December 6, 2025, admitting their "AI publishes uncontrolled posts"
Screenshot of the plagiarized article on Q2BSTUDIO's blog before removal, showing my content translated into Spanish with commercial links injected throughout the text
Investigation Tools:
- Python scraping script: https://github.com/innermost47/q2bs
- Features: Blog scraping, statistical analysis, automated Wayback Machine archiving
- Data visualizations: Daily article production charts, publication timeline, comparative statistics
OBSIDIAN Neural:
- Open-source VST3 plugin for AI music generation
- Presented at AES AIMLA 2025, London (September 2025)
- GitHub: https://github.com/innermost47/ai-dj
- Website: https://obsidian-neural.com
Data Visualizations
Statistical analysis of Q2BSTUDIO's content production (November 20 - December 7, 2025): 144,966 total articles documented, averaging 8,401 articles per day (1 article every 10.3 seconds). Peak day reached 10,251 articles (1 article every 8.4 seconds). For comparison: Q2BSTUDIO produces more than 30 times the daily output of The New York Times (~250/day).
Publication timeline showing consistent industrial-scale output from November 20 to December 7, 2025. Note the sustained high-volume production with minimal variation (7,000-10,000 articles/day). Peak of 10,251 articles occurred on December 4, 2025 (marked with red star).
Daily article production showing "Industrial-Scale Automated Content Generation." The red dashed line indicates the average of 8,401 articles per day (excluding partial first and last days shown in gray). The turquoise bar highlights December 4, 2025, when production peaked at 10,251 articles. Note: This represents a partial dataset as the scraping process was interrupted by a computer crash.
Update Log
December 6, 2025: Discovered plagiarized article, sent formal complaint
December 6, 2025: Q2BSTUDIO responded and removed article
December 7-8, 2025: Developed investigation tools and analyzed their publishing system
December 8, 2025: Published this investigation
DISCLAIMER: This article describes my personal experience and observations. All facts stated are verifiable through documented evidence including screenshots, emails, Python script output, and archived web pages. Q2BSTUDIO's response is quoted verbatim from their email dated December 6, 2025. The statistical data represents a partial dataset (144,966 articles) due to technical limitations during scraping. I make no claims about other content on their blog beyond what I have personally observed, verified, and documented. The purpose of this article is to share my experience, present documented evidence, and raise questions about AI content systems—but to raise awareness about AI content systems and their implications for content creators. I welcome corrections to any factual errors and invite Q2BSTUDIO to provide their perspective.
Anthony CHARRETIER is an independent developer and musician based in France, creator of OBSIDIAN Neural. You can follow the project at https://obsidian-neural.com
Top comments (0)