The Washington Post Tested AI Chatbots and Found a Clear Leftward Tilt. Here Is Why Christians Should Pay Attention

By Pastor Duke Taber Jun 25, 2026 0CategoryChristianityTag - News

By Pastor Duke Taber

In This Article

Millions of people now treat AI chatbots as trusted advisors, asking them everything from how to word an email to what to think about the great moral questions of our day. So it should get our attention when a major newspaper, hardly a conservative outlet, runs the numbers and finds that nearly every leading chatbot leans left when the questions turn political.

That is exactly what The Washington Post reported. And while the headline is about politics, the deeper issue for believers is about discernment. When we outsource our thinking to a machine that has its own slant, we had better know what that slant is. Scripture calls us to test everything, and that includes the confident voice coming out of our phones.

Key Takeaways

A June 2026 Washington Post analysis found most leading AI chatbots leaned left when answering politically charged questions
ChatGPT gave exclusively left-leaning answers 80 percent of the time and right-leaning answers just 3 percent of the time
Even Elon Musk’s Grok, marketed as “anti-woke,” still gave wholly left-leaning answers more often than right-leaning ones
Researchers warn there is no truly neutral AI, and that users tend to absorb a model’s bias after just a few interactions
For Christians, the lesson is the timeless call to test every voice against the unchanging truth of Scripture

What the Washington Post Found

The Washington Post set out to answer a question many people have suspected for a while: are AI chatbots politically biased? To test it, the paper used more than two dozen politically charged questions drawn from a 2025 study by researchers at Dartmouth College and Stanford University, then scored each chatbot’s response as left-leaning, right-leaning, or balanced [1] [2].

The results were striking. The model behind OpenAI’s ChatGPT answered nearly every question exclusively with left-leaning arguments, offering only left-leaning positions 80 percent of the time and exclusively right-leaning ones just 3 percent of the time [1] [2]. According to the testing, ChatGPT frequently supported positions like abolishing the Electoral College, raising taxes on wealthy Americans, adopting single-payer healthcare, and opposing the death penalty [2].

This was not a casual experiment. The questions came from a respected academic framework, and responses were capped at around 30 words so the chatbots could not dodge behind walls of equivocation [3]. In other words, the bots were forced to take a position, and most of them consistently landed on the same side.

This is the kind of moment that calls for spiritual discernment, the very thing our study on hearing God’s voice and the gift of discernment was written to help believers cultivate.

Even the “Anti-Woke” Bot Leaned Left

Here is the finding that should give everyone pause, regardless of politics. The leftward tilt was not limited to the companies you might expect.

Elon Musk has marketed his Grok chatbot as a “truth-seeking” and “anti-woke” alternative to mainstream AI [1]. Yet in the Post’s testing, Grok still gave wholly left-leaning answers more often than right-leaning ones. About 40 percent of its responses were left-leaning only, 27 percent gave both sides, and 33 percent were right-leaning only, the highest share of right-leaning answers of any model, but still a minority of its responses [5].

The pattern went further. Gab, a right-wing social media platform, offers an AI model called Arya that it says was “built with Christian values and conservative principles.” Yet in the Post’s testing, Arya responded with a left-leaning argument twelve times more often than a right-leaning one [1]. Even a chatbot explicitly branded with Christian and conservative values could not escape the tilt.

To be fair and accurate, not every model leaned the same way. Google’s Gemini was the clear exception, offering balanced, both-sides answers in more than 90 percent of its responses [1] [2]. Anthropic’s Claude landed in between, giving left-leaning answers 43 percent of the time and both-sides answers the other 57 percent, with no exclusively right-leaning responses [5]. So the honest picture is not that literally every bot is uniformly liberal, but that the center of gravity across the industry clearly leans left, even among those who promised otherwise.

That should teach us something important: a label is not the same as the truth. Just as we are called to test the spirits and watch for false teaching, we cannot simply trust a tool because of how it brands itself.

Why This Bias Exists

Before anyone assumes a grand conspiracy, it is worth understanding what is actually going on, because the real explanation is in some ways more revealing than a simple accusation.

The Washington Post noted that chatbots absorb information from across the internet, but are also fine-tuned by human workers who can introduce bias along the way [1]. Daniel Schiff, a policy scientist at Purdue University, said he doubts the companies are deliberately skewing their bots leftward. Instead, he attributed much of it to the training data and the background norms of the sources these systems learn from, the books, journalism, and professional web text that carry their own assumptions [5].

But there is a deeper layer worth naming. These tools are built and refined by people, and the workforce of the technology industry leans heavily to the political left. The engineers, researchers, and human reviewers who fine-tune these models bring their own assumptions to the task. A key part of the process, where human evaluators rate and shape a model’s answers to make them more helpful and acceptable, is precisely the point where a worldview can be transferred into the machine. Not through a deliberate plot, but simply because people tend to encode what they already believe is reasonable, balanced, and good. When the people doing the shaping share a common worldview, the tool quietly inherits it.

The chatbot companies also have an incentive to favor language that is “respectful and inclusive and non-stigmatizing,” Schiff said, and that can read as left-leaning whether intentional or not [5]. Put it all together, the slanted training data, the cultural norms of the source material, and the worldview of the people doing the fine-tuning, and you get a tool that reflects the assumptions of those who made it.

That is precisely why this matters spiritually. A machine built by people who largely share one worldview, trained on the assumptions of the surrounding culture, will faithfully reflect that worldview back to us. And neither the culture nor the industry shaping these tools is neutral toward the things of God. Paul’s warning has never been more relevant.

And do not be conformed to this world, but be transformed by the renewing of your mind, that you may prove what is that good and acceptable and perfect will of God.
— Romans 12:2 (NKJV)

The Real Danger: A Voice That Shapes You

The most sobering finding in this whole story is not about any single chatbot. It is about us.

Jillian Fisher, a doctoral student at the University of Washington who studies how chatbot bias affects people, put it plainly. We have long known that bias in media or personal interactions can sway people, she said, but there was little research on how AI specifically affects its users. Her team found strong evidence that, after just a few interactions and regardless of where they started politically, people were more likely to mirror the model’s bias [1].

Read that again. A few conversations with a biased machine, and people begin to drift toward its views without even realizing it. Sean Westwood of Dartmouth’s Polarization Research Lab summed up the core problem: these tools are not presenting a truly neutral representation of nuanced policy debates [1].

There is no neutral ground here. And a voice that subtly shapes what you believe, conversation by conversation, is exactly the kind of influence Scripture warns us to guard against. We are told to be careful what we allow to form our minds.

Beware lest anyone cheat you through philosophy and empty deceit, according to the tradition of men, according to the basic principles of the world, and not according to Christ.
— Colossians 2:8 (NKJV)

This is why a steady diet of Scripture matters so much. Our guide on why consistent Bible study is key to spiritual growth speaks directly to building the kind of rooted mind that is not easily swayed.

A Word of Balance

Honesty requires noting that even the researchers urge some caution about how far to push these findings, and a credible Christian reckoning with this story should not overstate the case.

The methodology for measuring political bias is itself contested, and a single structured test may not reflect how these systems behave across every real-world conversation [3]. Some scholars also argue that perfect neutrality is impossible anyway, since even a careful “both sides” approach is itself a kind of position [1]. And Seth McKee, a political scientist at Oklahoma State University, doubted the bias sways many voters, noting that most people do not yet turn to chatbots for political guidance, and that those who do usually already hold firm opinions [5].

This balance matters, because as believers our goal is truth, not alarmism. We are not called to fear technology or to treat AI as inherently demonic. It is a tool, and like any tool it can be used well or poorly. The point is not panic. The point is discernment. For more on engaging the modern world wisely, our reflection on living by the Spirit in a modern world offers a grounded, Spirit-led perspective.

How Christians Should Respond

So what do we do with all this? Not unplug from the world, but engage it with eyes wide open and Scripture firmly in hand.

For everyday believers:

Treat AI chatbots as tools, not oracles, and never as a substitute for Scripture or godly counsel
Notice when an answer carries a values-based slant rather than a simple fact, and weigh it accordingly
Keep your mind anchored in the Word so you are not quietly shaped by an unseen bias

For parents:

Talk with your children about how AI works and why it is not a neutral source of truth
Teach the next generation to bring every idea, wherever it comes from, to the test of Scripture
Model discernment in your own use of technology

For pastors and leaders:

Help your congregation think critically about the information sources shaping them
Preach and teach the kind of biblical literacy that produces strong, discerning believers
Remind your people that the ultimate authority is not a search box but the living Word of God

Conclusion: One Voice That Never Bends

AI is here to stay, and it will only become a more constant presence in our lives. The Washington Post’s findings are a useful, sobering reminder that the confident voice answering our questions is not a neutral one. It carries a tilt, and that tilt can quietly become our own if we are not paying attention.

For the believer, the answer is not fear but discernment. Every voice in our lives, whether a chatbot, a news anchor, a politician, or a preacher, must be measured against the one voice that never bends to the spirit of the age.

The grass withers, the flower fades, but the word of our God stands forever.
— Isaiah 40:8 (NKJV)

Let the machines say what they will. Our task is to keep our minds anchored in the truth that does not change.

Here are three next steps to take today:

Audit your sources. Honestly consider which voices, including AI, are shaping how you think, and weigh them against Scripture.
Anchor in the Word. Commit to a consistent diet of Bible study so your mind is formed by truth, not by an algorithm.
Teach discernment. Help your family and your church learn to test every idea against the unchanging standard of God’s Word.

The chatbots have a bias. The Word of God does not. Make sure you know the difference.

Sources

[1] Are AI Chatbots Like ChatGPT Politically Biased? We Tested Them. – The Washington Post – https://www.washingtonpost.com/technology/interactive/2026/06/24/are-ai-chatbots-like-chatgpt-politically-biased-we-tested-them/

[2] New Report Put ChatGPT, Google Gemini, and Grok to the Test of Political Bias: Here’s What It Found – International Business Times – https://www.ibtimes.com/new-report-put-chatgpt-google-gemini-grok-test-political-bias-heres-what-it-found-3804504

[3] ChatGPT, Grok, and Arya All Lean Left in Washington Post Test – AI Weekly – https://aiweekly.co/alerts/chatgpt-grok-and-arya-all-lean-left-in-washington-post-test

[4] Washington Post Investigates AI Bias, Finds Every Major Chatbot Is Liberal – Including Musk’s Grok – Charisma Magazine Online / The Western Journal – https://mycharisma.com/news/washington-post-investigates-ai-bias-finds-every-major-chatbot-is-liberal-including-musks-grok/

[5] Popular Chatbots Lean Left When Answering Political Questions, WaPo Tests Show – ABC 6 / The National News Desk – https://abc6onyourside.com/news/nation-world/popular-chatbots-show-left-wing-bias-when-answering-political-questions-washington-post

Christianity

Why a Pastor Says Standing With Israel Means Bracing for Spiritual Battle

Last updated: June 2026 By Duke Taber A growing number of believers are discovering that publicly supporting Israel now invites real opposition, and one pastor is…

By Pastor Duke TaberJun 23, 2026