Uncategorized
PoliticIt AI & Editorial Transparency Disclosure
Last Updated: January 10, 2026
PoliticIt is committed to transparency, constitutional principles, accurate political discourse, and public trust. We use artificial intelligence as an editorial and production tool, never as a replacement for human judgment, authorship, or accountability. This policy exceeds Utah’s legal requirements and reflects best practices for media in the AI era.
1. Scope of AI Use at PoliticIt
We employ a combination of commercial services and private infrastructure.
Commercial Services:
- Claude (Anthropic) for deep legal nuance
- ChatGPT (OpenAI) for clarity accessible to citizens
- Grok (xAI) for cross-domain inference
- Gemini (Google) for large-scale pattern detection
- Suno for music generation
Private Infrastructure:
- Local Retrieval-Augmented Generation (RAG) system on NVIDIA DGX Spark hardware (GB10 Grace Blackwell Superchip with 128GB unified memory), enabling private, offline processing of large verified corpora without transmitting data to external AI providers. This system supports large, state-of-the-art language models.
- Open-source speech synthesis and music tools
Functions Supported:
- Research synthesis, source analysis, and initial drafting
- Podcast transcript generation and audio editing
- Background music creation
- Content formatting, grammar, structure, and readability improvements
These systems serve as editorial assistants and production tools, comparable to research assistants, transcription software, editing programs, and audio equipment. They do not operate autonomously.
AI does not:
- Decide editorial positions, endorsements, or factual conclusions
- Publish content without human review and approval
- Create deceptive synthetic media (deepfakes or altered images, audio, or video intended to mislead about events, statements, or identity)
All published content is extensively reviewed, substantively edited, fact-checked against primary sources, and approved by a named human author, editor, or producer who bears ultimate responsibility.
2. Human Accountability and Editorial Process
Content reflects the views and final judgment of human creators. AI-assisted material undergoes the same rigorous standards for accuracy, fairness, sourcing, and compliance.
Key Review Steps:
- Factual verification from primary sources
- Alignment with constitutional governance principles
- Substantive editing for coherence and integrity
- Multi-model cross-verification, with human evaluation of discrepancies, to check for consistency and catch errors
- Adversarial red-teaming: prompting models to identify weaknesses, challenge assumptions, and stress-test reasoning before publication
- Visual content review to ensure images are authentic, appropriately sourced, or clearly disclosed as AI-generated or illustrative
- Manual citation verification
- Final human approval
Material assistance includes substantive contributions like drafting, restructuring, transcripts, or production. Minor tasks like spell-checking do not trigger disclosures.
Our approach draws on decades of AI expertise, from early neural networks and genetic algorithms through parallel processing and founding pre-ML AI systems. This background helps us guard against pitfalls like hallucinations and bias.
3. Compliance with Utah Law
We comply with Utah Code § 20A-11-1104, which addresses synthetic media in paid election-influencing communications.
While PoliticIt’s editorial and podcast content does not typically fall under this statute’s “paid communications” definition (which targets candidate and PAC advertisements), we voluntarily apply disclosure standards that meet or exceed those required for paid synthetic media to maintain public trust and prevent any deceptive use of AI-generated content.
Our commitments:
- No production or publication of deceptive synthetic audio, video, images, or text impersonating individuals or fabricating events or statements without required conspicuous disclosures
- No misleading AI-generated images regarding authenticity of persons, events, or documents without disclosure
- When applicable, prominent notice such as: “This contains content generated or altered by artificial intelligence”
Creative Exceptions: AI-generated background music, illustrative graphics, or clearly labeled satire or parody that cannot reasonably be mistaken for authentic reporting or real-world events requires no per-instance disclosure beyond this policy.
Broader generative AI rules under the Utah Artificial Intelligence Policy Act (as amended in 2025) apply primarily to consumer and high-risk interactions. Our editorial focus prioritizes election-specific synthetic media compliance.
4. Permitted and Prohibited Assistance
Permitted (with human oversight):
- Research, drafting, and transcripts
- Audio editing and production
- Background music composition
- Readability and structure improvements
- Formatting, translation, and accessibility features
Prohibited:
- Publishing without extensive human review and fact-checking
- Deceptive impersonation or fabrication
- Undisclosed synthetic media of candidates or officials in false election contexts
- Obscuring political messaging origins
5. Disclosure Implementation
Site-Wide Footer:
All PoliticIt content includes a universal disclosure footer linking to this comprehensive policy. This site-wide approach ensures consistent transparency across all articles, podcasts, and content.
Footer text:
“AI DISCLOSURE: PoliticIt uses artificial intelligence tools to assist with research, drafting, transcription, and content production. All content is extensively reviewed, fact-checked, and approved by named human editors who bear full responsibility for published material. AI is a tool, not a speaker. Read our full AI & Editorial Transparency Disclosure: politicit.com/ai-disclosure”
Optional Article-Specific Context:
For content with particularly extensive AI assistance, novel AI applications, or where additional context would be helpful, we may include supplementary notes within the article itself.
6. Data Privacy and Infrastructure
Private local RAG systems operate without internet access and without external training data, minimizing third-party sharing. When using commercial services, we avoid inputting confidential or non-public data and retain full editorial control.
7. Commitment to Transparency
We support targeted disclosure laws that inform audiences and prevent deception without overreach, prior restraint, or viewpoint discrimination.
AI is a tool, not a speaker. Its use preserves First Amendment protections and human accountability.
8. Why This Approach Matters
PoliticIt’s AI disclosure policy reflects three core commitments.
Transparency: Citizens have a right to know when and how AI is used in political content, even when not legally required.
Accountability: AI tools enhance human analysis. They don’t replace human judgment, sourcing, or responsibility. Every claim we publish has been verified by a human editor.
Access: By combining commercial AI services with sophisticated private infrastructure, we provide constitutional and policy analysis at a depth previously available only to well-funded institutions. This helps level the playing field for informed civic participation, enabling citizens to engage with complex legal and political questions previously accessible only through expensive consultants or large institutional research departments.
9. Feedback and Updates
Contact: contact@politicit.com
This policy will be reviewed at least annually and updated within 30 days of any material change to Utah Code § 20A-11-1104 or related election disclosure statutes, or as technology and practices evolve.
Summary
PoliticIt uses advanced AI, including local high-capacity infrastructure and commercial tools, as controlled editorial and production aids under strict human oversight. We do not deceive, impersonate, or replace humans. All content is human-reviewed, fact-checked, and accountable to named editors and authors.
Published in compliance with Utah law, including Utah Code § 20A-11-1104.
