Disclaimer
Last updated: 17 May 2026.
What BiasFeed is
BiasFeed is an independent, experimental project that uses artificial intelligence to illustrate how the same set of facts can be framed to produce very different narratives. Its purpose is to foster media literacy by making the mechanics of political framing visible. The site is provided for illustrative and educational use, and should not be treated as a news service or a definitive source of truth.
How each article is structured
Every article on this site presents the same underlying story in three forms:
- Facts: a list of factual claims extracted from the original source material by an AI model. This is the only part of an article that attempts to state facts, and even then it inherits any limitations of the source.
- Centrist Version: an AI-generated retelling that aims for neutral language.
- Left-Biased Version and Right-Biased Version: AI-generated retellings that deliberately apply rhetorical framing typical of each end of the political spectrum.
The Left-Biased and Right-Biased versions are commentary, not reporting
The Left-Biased and Right-Biased versions on this site are opinion and commentary, not statements of fact. They are produced by a third-party large language model not controlled by the site operator, and are intentionally written to exaggerate the rhetorical posture of each side. They use adjectives, framing, emotive language, and sometimes sharply pointed characterisations of named people, organisations, and policies, as a technique to demonstrate how the same facts can be spun.
Any characterisation of a named individual, company, government, party, or public figure in those versions is part of the rhetorical framing and should not be read as a factual claim about that person or entity's conduct, motives, or character. If you want to know what actually happened, read the Facts tab and follow the source links to primary reporting.
Not the views of the site operator
The opinions expressed in the Centrist, Left-Biased, and Right-Biased versions are generated by an AI language model. They do not represent the views, opinions, or political positions of BiasFeed, its operator, or anyone associated with the project. Publishing a Left-Biased or Right-Biased version is not an endorsement of either side; the whole point of the site is to put both side by side so the framing becomes visible.
Do not rely on AI-generated content as news
AI-generated text can contain factual errors, omissions, misinterpretations, hallucinated detail, and incorrect attribution, even when the underlying source is sound. The content on this site is not a substitute for primary reporting. Before forming a view, sharing, or acting on anything you read here, follow the source links and read the original journalism.
Public Beta and work in progress
BiasFeed is in a public Beta phase. The AI prompts and pipelines that produce the Facts list and the rewritten versions are under continuous development. The quality, accuracy, tone, and nuance of the output may vary, sometimes significantly, while the system is refined.
Automation and editorial review
The content generation process is largely automated. An AI model extracts facts from a source article, and subsequent AI models rewrite those facts into different versions. Due to the automated nature of this pipeline, not every article is reviewed by a human editor before publication. Errors in fact extraction, interpretation, or rewriting are possible and should be expected from time to time.
Corrections, complaints, and takedown requests
If you believe an article on this site contains a clear factual error, misrepresents a named person or organisation, or causes you concern for any other reason, please contact us at [email protected]. Tell us which article (a URL is best) and what specifically is wrong.
Given the experimental nature of the site and its limited resources, our standard remedy is the permanent removal of the affected article. This ensures that potentially flawed or contentious content is taken down quickly. Minor edits may be made in some cases, but removal is the default. We aim to action genuine requests promptly.
Limitation of liability
The content on BiasFeed is provided "as is", for educational and illustrative purposes, without warranty of any kind. To the maximum extent permitted by law, the operator of BiasFeed assumes no liability for any errors, omissions, inaccuracies, or consequences arising from reliance on the content of this website. Use of this site is at your own risk.
Governing law
BiasFeed is operated from Australia. This disclaimer and any matter arising from your use of the site are governed by the laws of Australia.
Changes to this disclaimer
We may update this disclaimer from time to time as the site evolves. The "Last updated" date at the top of this page reflects the date of the most recent change.