Editorial Policy & Standards

1. The Foundation of Trust

Trust is the most valuable asset of any research institution. In the context of AI software—where capabilities are often overstated and risks are obscured—unbiased, accurate reporting is critical. WhichAIPick adheres to a rigorous set of editorial standards designed to protect our readers and ensure the integrity of our data. We model our code of ethics on frameworks such as the SPJ Code of Ethics, adapted for the algorithmic age.

These policies apply to every contributor, editor, and analyst associated with WhichAIPick. There are no exceptions.

2. Editorial Independence Guidelines

The separation between our "Church" (Editorial/Research) and "State" (Business/Partnerships) is absolute. We safeguard this independence through specific protocols:

The Firewall Protocol

Our Editorial Board has final say on all content. Our Business Development team has zero access to our Content Management System (CMS) for the purpose of editing reviews. They cannot draft, edit, or delete editorial content.

Embargoes and NDAs

We frequently receive pre-release access to AI models under Non-Disclosure Agreements (NDAs) or Embargoes. We honor these legal agreements to allow us to test tools before they launch publically. However, we do not sign NDAs that restrict our ability to publish a negative review. If a tool is flawed, we reserve the right to say so the moment the embargo lifts.

Pay-to-Play Zero-Tolerance

To be explicitly clear:

• We do not sell scores. A vendor cannot purchase a "90/100" rating.
• We do not sell awards. "Top Pick" badges are awarded solely on merit.
• We do not sell takedowns. A vendor cannot pay to have a negative review removed.

3. What We Do / What We Don't Do

In plain English, here are the boundaries of our coverage:

4. Fact-Checking and Verification

In the era of AI hallucination, verification is paramount. We hold our content to an academic standard of accuracy. See our Review Methodology for the full testing protocol.

Primary Source Requirement

We do not rely on third-party news reports or "rumors" on Twitter/X. All technical specifications (e.g., context window size, parameter count, pricing) must be verified via:

• Official Vendor Documentation (primary).
• Direct API Testing / Logs (primary).
• Peer-Reviewed ArXiv Papers (primary).
• Direct Interview with Founders/CTOs (secondary).

The "Two-Eye" Rule

No review requires only one person's approval. Every published piece undergoes a "Two-Eye" review process:
1. The Analyst: Conducts the testing and drafts the copy.
2. The Managing Editor: Verifies the claims, checks the tone, and ensures adherence to the style guide.

5. Conflict of Interest Controls

We operate in a small industry. Connections between analysts and founders are inevitable. To manage this, we enforce strict disclosure rules:

• Financial Disclosure: Staff members are prohibited from reviewing products from companies in which they hold direct equity or stock options.
• Relationship Disclosure: Staff must recuse themselves from reviewing products built by immediate family members or close personal friends.
• Gift Policy: We do not accept gifts, travel, or accommodation from vendors. We pay our own way to conferences. We accept "Review Keys" (free access to software) strictly for the purpose of evaluation, with the understanding that this does not garnish favor.

6. Sourcing and Attribution

We respect intellectual property. When we cite data, benchmarks, or news broke by another publication, we provide clear, visible attribution.

We do not aggregate content. If we are reporting on a new model release, we add value through our own testing rather than simply rewriting a press release.

7. Diversity and Inclusion in AI

We recognize that the AI industry currently suffers from a lack of diversity. Algorithmic bias is a direct result of homogenous training data and homogenous development teams. WhichAIPick is committed to:

• Highlighting Diverse Founders: We actively seek out tools built by underrepresented groups.
• Testing for Bias: Our Review Methodology includes specific tests for gender and racial bias in generative outputs. We call out tools that fail these safety checks.

8. Use of Generative AI in Our Newsroom

It would be hypocritical for an AI research site to ban AI. However, we use it responsibly as an assistive technology, not a replacement for journalism. We follow the Academy Guidelines for responsible AI interaction.

Acceptable Uses:

• Summarizing massive PDF whitepapers to extract key technical stats.
• Generating initial code snippets to test API functionality.
• Fixing grammar and spelling errors (copyediting).

Prohibited Uses:

• Asking an LLM to "Write a review of [Tool X]" (This leads to hallucination and generic fluff).
• Generating photorealistic images to represent real events (Misinformation).
• Publishing AI output without human verification.

9. Corrections and Retractions

We believe that admitting mistakes builds trust. If we get something wrong, we fix it—fast and visibly.

• Corrections: For factual errors (dates, prices, specs), we fix the text and add a note at the bottom of the piece: "Correction [Date]: An earlier version of this article incorrectly stated..."
• Retractions: If a review is fundamentally flawed (e.g., our testing methodology was misapplied), we will retract the piece entirely, leaving a transparency note explaining why.

See our Corrections Policy for the full protocol.

If you disagree with a finding, read How to Challenge Our Findings.

10. Comment Moderation

We encourage community discussion but maintain a high bar for civility. We reserve the right to delete comments that contain:

• Hate speech or harassment.
• Spam or self-promotion.
• Verifiably false disinformation regarding AI safety.

11. Limitations of Our Policy

We cannot police the entire internet. Our policy applies to content hosted on whichaipick.com. We are not responsible for the content of third-party sites we link to.

Frequently Asked Questions

Related Resources

• Review Methodology
• Submit a Correction
• Affiliate Disclosure
• About Us
• Compare Tools