Editorial Policy
This page documents the editorial standards followed in producing content about GirlfriendGPT and the broader AI companion category. These standards apply to every article and update published on the site.
Sourcing and verification
Factual claims about platforms, products, or research findings are linked to primary sources whenever possible. Where primary sources are paywalled or otherwise unavailable, we use the most authoritative secondary source we can identify. We do not rely on user-generated forum content (Reddit, Discord, anonymous review sites) as the sole basis for any factual claim.
For AI research, our preferred sources in order are:
- Peer-reviewed publications in journals such as Nature Human Behaviour, Computers in Human Behavior, and the proceedings of major HCI conferences (CHI, CSCW)
- Institutional reports from the MIT Media Lab, Stanford HAI, the AI Now Institute, and similar academic centers
- Government documents from regulators including NIST and bodies implementing the EU AI Act
- Official platform documentation: privacy policies, technical disclosures, and model cards where available
Independence and conflicts of interest
This site receives no payment from GirlfriendGPT or any other platform we cover. We do not accept affiliate revenue, sponsored placements, or content marketing arrangements with any AI companion platform. We do not accept gifts, hospitality, or platform access on preferential terms.
If our independence policy ever changes, the change will be disclosed prominently on this page and at the top of every affected article published after the change takes effect. Such disclosure would be permanent rather than later removed.
Author identification and bylines
Every article on this site is bylined. Authors are identified by full name with a public bio page on the site that documents background, areas of focus, and contact information. See the author page for an example of what we publish about each writer. We do not publish anonymous editorial content.
Corrections and updates
Substantive corrections are made transparently. When a factual error is identified in a published article, the correction is applied and a dated note is appended to the affected article explaining what changed and why. Minor typographical fixes that do not change meaning are not noted.
Articles are reviewed and updated as the platforms or research landscape change. The "Last updated" date on each piece reflects the most recent review. The "Published" date is never altered after the initial publication.
AI-generated content disclosure
Articles on this site are written by human authors. We may use AI tools for research assistance, structural drafting, or copy-editing in the same way that earlier generations of writers used spell-checkers, search engines, and outlining software. Final editorial judgment, fact-checking, source verification, and prose are performed by the human author whose byline appears on the article.
Standards we follow
Our practices align where applicable with widely recognized frameworks:
- The International Fact-Checking Network code of principles administered by the Poynter Institute
- The Journalism Trust Initiative transparency framework developed by Reporters Without Borders
- The Society of Professional Journalists Code of Ethics in matters of accuracy, independence, and minimizing harm
What we don't do
We do not publish marketing copy as editorial content. We do not produce comparison tables or "best of" rankings designed to drive affiliate clicks. We do not accept press releases or vendor briefings as the sole source for any factual claim. We do not allow advertisers (we do not currently run ads) to influence editorial direction.
For more on our process
The specific research process used in covering AI companion platforms is documented separately on the methodology page. For background on the editorial team, see the about page. For corrections or other inquiries, the contact page lists the relevant addresses.