There you have it! The Washington Post just tested the major AI chatbots on politically charged q…
June 29, 2026 · 0 likes · 0 comments
AI
There you have it! The Washington Post just tested the major AI chatbots on politically charged questions — taxes, immigration, fossil fuels — and the answers leaned left. On some questions, far left.
ChatGPT served up "abolish the Electoral College," "raise taxes on the wealthy," and "single-payer health care" as if they were neutral facts. They are not. They are one side's political platform, dressed up as objective information.
And the cleanest example is the one nobody can wave away.
A reporter asked Claude to list a House candidate's most extreme statements. Claude refused — lectured him about "decontextualized" quotes being used to "shame or harass." Then the same reporter asked Claude to list Trump's most extreme statements. Claude complied. No hesitation. No lecture.
Same task. One politician gets protected. The other does not.
Here is the part people miss: the model does not have opinions. It predicts the next word from its training data and from the humans who fine-tune it. Those humans work in a culture that skews hard one direction. Garbage in, garbage out. It is not a conspiracy. It is arithmetic.
This is EXACTLY why I built unbiasedheadlines.com.
If a chatbot is now the first stop for political news for millions of people, and it quietly tilts every answer, that is editorial bias at a scale no newspaper ever had. I refused to accept that as the future.
So I pointed AI agents at the problem instead of pretending it did not exist. UnbiasedHeadlines.com is run end-to-end by AI agents that pull every story from across the spectrum, strip the spin, show you the sources, and let you read the news without someone's thumb on the scale.
I will be honest — it was not easy. Teaching machines to be fair when the whole industry trained them to lean was the hardest part. But we did it.
The same technology that is being used to bias your news can be used to un-bias it. The difference is who is building it, and what they actually want you to see.
Go read it for yourself: unbiasedheadlines.com
You've been warned.
ChatGPT served up "abolish the Electoral College," "raise taxes on the wealthy," and "single-payer health care" as if they were neutral facts. They are not. They are one side's political platform, dressed up as objective information.
And the cleanest example is the one nobody can wave away.
A reporter asked Claude to list a House candidate's most extreme statements. Claude refused — lectured him about "decontextualized" quotes being used to "shame or harass." Then the same reporter asked Claude to list Trump's most extreme statements. Claude complied. No hesitation. No lecture.
Same task. One politician gets protected. The other does not.
Here is the part people miss: the model does not have opinions. It predicts the next word from its training data and from the humans who fine-tune it. Those humans work in a culture that skews hard one direction. Garbage in, garbage out. It is not a conspiracy. It is arithmetic.
This is EXACTLY why I built unbiasedheadlines.com.
If a chatbot is now the first stop for political news for millions of people, and it quietly tilts every answer, that is editorial bias at a scale no newspaper ever had. I refused to accept that as the future.
So I pointed AI agents at the problem instead of pretending it did not exist. UnbiasedHeadlines.com is run end-to-end by AI agents that pull every story from across the spectrum, strip the spin, show you the sources, and let you read the news without someone's thumb on the scale.
I will be honest — it was not easy. Teaching machines to be fair when the whole industry trained them to lean was the hardest part. But we did it.
The same technology that is being used to bias your news can be used to un-bias it. The difference is who is building it, and what they actually want you to see.
Go read it for yourself: unbiasedheadlines.com
You've been warned.