Bible App CEO: AI chatbots misquote Scripture up to 60% of the time — and it's a generational threat
YouVersion's Bobby Gruenewald is quietly pressuring OpenAI, Anthropic, and Google to fix how their models handle sacred text — while refusing to add a chatbot to the world's most-downloaded Bible app.
The most-downloaded Bible app in the world has more than a billion installs, operates in hundreds of languages, and is, by its CEO's own account, sometimes the last thing a suicidal person reaches for before deciding whether to live. Bobby Gruenewald, founder and CEO of YouVersion, has spent years thinking about what that responsibility means — and right now, his biggest concern is not a rival app or a funding gap. It is ChatGPT, Claude, and Google Gemini getting Scripture wrong.
Gruenewald has been making the case publicly and privately that AI's misquotation of the Bible is not a minor technical glitch but a slow-moving threat to the integrity of a text that took centuries of careful translation to produce. Speaking with Christianity Today in New York, where he was unveiling a Times Square ad campaign for Global Bible Month, he laid out the scale of the problem in plain numbers.
The best model with the best performance, with the most popular versions of the Bible that are most indexed, misquotes Scripture at least 15% of the time. Some of them as much as 60% of the time.— Bobby Gruenewald, founder and CEO of YouVersion
The errors are not always dramatic. Some involve commas, minor wording shifts, or subtle paraphrasing. But Gruenewald argues that in Bible translation, small changes are not small. 'Every word and punctuation is meaningful to Scripture translation,' he said. If AI models quietly alter verses at scale — and hundreds of millions of users treat those outputs as authoritative — the cumulative effect could degrade the translations themselves over a generation.
Misquoting the Bible becomes a really big generational problem. If all the work and care that goes into translating the Bible correctly just erodes because the translations become modified in a subtle way over a generation, then I feel like we've failed in our responsibility to steward this text.— Bobby Gruenewald, founder and CEO of YouVersion
The root of the problem, Gruenewald says, is how large language models work. They train on vast swaths of the internet, which makes them broadly capable but structurally unpredictable. Unlike conventional software, where a wrong answer is a bug that can be patched, AI is nondeterministic — it has variability built in. 'We're used to having software where, when it doesn't answer correctly, it's a bug and you fix the bug,' he told Christianity Today. 'That's not the same way that this technology works.'
Unfortunately people view the technology right now as though it's an oracle that you can ask any question about anything, and it's supposedly smarter than you. Human beings have generally associated fluency with intelligence. So what you have is a technology that, by design, is super fluent.— Bobby Gruenewald, founder and CEO of YouVersion
That fluency is precisely what makes the errors hard to catch. Users who do not already know a verse have no way of knowing whether the version an AI returns is accurate, paraphrased, or fabricated. New data cited in the Christianity Today interview found that half of the Christian questions sent to ChatGPT are about interpreting the Bible — meaning the exposure is not marginal.
YouVersion has been quoting the Bible accurately, by direct quotation, for more than 18 years across its apps. The organization is now working on a benchmark to measure how well the major AI labs handle Scripture, and has been engaging privately with OpenAI, Anthropic, and Google to push for improvement. Gruenewald said YouVersion would help the labs gain access to reliable biblical texts if they could demonstrate consistent accuracy.
That private pressure campaign has not yet produced a public-facing AI feature inside the Bible App — and Gruenewald says that is deliberate. YouVersion built a chatbot prototype four or five years ago and chose not to launch it. The decision came down to two concerns: accuracy and safety.
In our case there is an implied authority that we don't actually even intend to have. But it does happen because people attribute something coming from the Bible App as something that's coming from God himself. And that's not proper theology. But we know that that happens, and so we have to be really thoughtful and careful.— Bobby Gruenewald, founder and CEO of YouVersion
The safety concern is concrete. Gruenewald said people in suicidal crises download the app regularly — it functions, for some users, as a last resort. An open-ended chatbot asking 'What does the Bible say about ___?' would, in his view, become an open invitation for people in desperate situations to seek answers from a system that cannot recognize the gravity of what it is being asked. His position: if a conversation turns toward suicide, the chatbot should stop entirely and route the user to a trained human.
Responsibly, you would want a human that had training and education and understanding to be able to navigate some of the complex situations they're in.— Bobby Gruenewald, founder and CEO of YouVersion
YouVersion does use AI — internally, for tasks like accelerating coding and improving workflow. Gruenewald describes himself as an early adopter of the technology. His objection is not to AI in principle but to deploying it in a context where errors carry spiritual weight and users are unlikely to know when they are being misled.
If we ever do fully adapt AI, it will be because we feel very confident that it can be done safely and be done with a level of accuracy and integrity.— Bobby Gruenewald, founder and CEO of YouVersion
When it comes to answering life's most important questions and trying to give direction from God's Word, we need it to be better in order to rely on it.— Bobby Gruenewald, founder and CEO of YouVersion
The broader Christian technology landscape is moving in different directions simultaneously. According to israelitestudies.com, Axios reported in 2025 that congregations in the United States have begun using AI to draft sermons, create devotional materials, and power prayer apps, with some platforms allowing users to chat with biblical characters. Faith-oriented technology company Gloo, according to a Reuters report cited by israelitestudies.com, launched efforts in 2025 to evaluate AI systems against values important to Christian communities, including theological integrity. Gruenewald's position — neither full rejection nor full embrace — reflects a fault line running through ministry leadership more broadly.
For now, the benchmark Gruenewald's team is developing would give the AI labs a measurable standard to meet on Scripture accuracy. Whether that standard moves the frontier models — which are built for general use and not specifically for religious texts — remains to be seen. What is clear is that the question will not go away: younger generations are already turning to chatbots before they turn to clergy, and the apps they use to read the Bible are under pressure to keep up.
Why it matters — With a billion-plus users treating the Bible App as an authoritative source, any AI feature that silently corrupts Scripture — even by a comma — could reshape how a generation reads and remembers the text, and YouVersion's public pressure on the major AI labs is one of the few organized efforts to set a measurable standard before that happens.
⚠ Not yet confirmed
- Half of the Christian questions sent to ChatGPT are about interpreting the Bible.
- incorrect source_host attributions on multiple quotes (changed to match christianitytoday.com or israelitestudies.com as appropriate)
Reported by christianitytoday.com, israelitestudies.com