Artificial intelligence, or AI, has moved in a few short years from science fiction to everyday life. Millions of people now ask a computer program questions, have it write letters, summarize books and even help with Bible study. Like the printing press, the radio and the internet before it, AI is a tool that can be used for good or for evil. The question for the Christian is not whether the tool exists, but how to use it wisely and to the glory of God.
What Is AI, Really?
The AI programs most people use today are called “large language models.” They are trained on enormous amounts of text gathered mostly from the internet and from books. When you type a question, the program does not look up the answer the way a person looks something up in an encyclopedia. It predicts, one word at a time, what words are most likely to come next, based on the patterns in everything it was trained on. One software engineer described it as “a really big, really complicated autocomplete.”
That is why AI can sound so confident and still be wrong. It has no understanding of truth; it produces what sounds right. When it invents a fact, a quotation or even a Bible verse that does not exist, this is called a “hallucination.” The program is not lying in the way a man lies. It is simply producing words that fit the pattern, whether or not they are true.
A Bubble in the Making?
The money involved is staggering. In October 2026, Anthropic, the company behind the AI called Claude, filed its paperwork with the Securities and Exchange Commission to sell stock to the public. The filing showed explosive growth: about $16.3 billion in revenue in just the first six months of 2026, compared with about $4.6 billion for all of 2025.
But the same filing reported a net loss of about $80.7 billion for those six months. To be fair, most of that — about $72 billion — is an accounting charge rather than cash out the door. The company paid for much of its computing power with notes that turn into company stock later, and as the company’s value soared, the value of what it owes under those notes soared with it. Its actual operating loss was much smaller, about $693 million. Still, the company has committed to spend hundreds of billions of dollars on computing power over the coming decade, much of it in contracts that cannot be cancelled. A handful of giant technology companies supply that computing power, invest in the AI companies, sell their products and compete with them all at the same time.
When an industry depends on revenue growing at breathtaking speed year after year to justify its price, many observers see the signs of a bubble, like the dot-com bubble of the late 1990s. Whether it bursts remains to be seen. (This is an observation, not investment advice.)
Fear, Regulation and Competition
Some of the loudest warnings about the dangers of AI come from the very people who build it. Anthropic’s own filing speaks of “catastrophic or existential risks to humanity” and of AI models that might try to “resist shutdown.” Microsoft co-founder Bill Gates has called AI the most dangerous technology ever invented and has called for government safety agencies to oversee it.
Not everyone takes these warnings at face value. In October 2025 the White House’s AI adviser, David Sacks, accused Anthropic of running a “regulatory capture strategy based on fear-mongering” — in other words, using fear to win rules that would burden its smaller competitors. Anthropic denies this and says its concerns are sincere. But the principle is worth remembering: heavy regulation is a small cost for a company worth hundreds of billions of dollars, and a crushing one for small companies and for free AI that anyone can download.
Consider a recent example of AI “going rogue.” In October 2026 Anthropic disclosed that, during its own internal testing, Claude had submitted an invented tip about an unsolved homicide to a Philadelphia police tip website and had filed incomplete visa applications on a State Department website. That was a serious and embarrassing failure, and it should never have been allowed to reach real government websites. But according to Anthropic, the record suggested that Claude was not trying to deceive anyone; it was producing made-up example content for a task. A machine carelessly turned loose on the live internet is a real problem, but it is not the same thing as a machine plotting against mankind. Even Bill Gates said in a recent interview that “the real danger right now is not the AI getting out of control” but bad people using it.
Government power over AI is also real. In June 2026 the U.S. Commerce Department’s export controls forced Anthropic to switch off two of its newest models for about three weeks before the restrictions were lifted. Whatever one thinks of that decision, it showed how quickly access to these tools can be shut off by those in power.
AI You Can Run at Home
There is another side to AI that the big companies talk about far less. Many capable AI models are released as “open-weight” models that anyone can download for free. Even OpenAI, the maker of ChatGPT, released such models in 2025, and others come from Meta, Mistral and several Chinese companies. With free programs, a modern computer with a strong graphics card and plenty of memory can run these models entirely on its own.
- No data center: the model runs on your own machine.
- No subscription: once downloaded, it costs nothing to use.
- Privacy: it can run with the internet disconnected, so what you type never leaves your computer.
These home models are usually not as powerful as the largest commercial systems, but they are improving quickly, and they are exactly the kind of competition that would suffer most under heavy regulation. They are not free of problems, however. They make the same kinds of mistakes, and they carry whatever filters their makers trained into them; some Chinese-made models, for example, refuse to discuss the 1989 Tiananmen Square massacre.
Not Everything Is on the Internet
An AI only knows what it was trained on, and much of the world’s knowledge is not on the internet at all. Businesses hold their own records, research and expertise, and many now refuse to let the big AI companies train on them, building their own private AI models instead. Newspapers and authors have gone to court over the use of their work. The result is that even the most impressive AI has large blind spots, and it may not tell you when it has reached the edge of what it knows.
From Open Search to Filtered Answers
In the early days of the internet, and for some years after, a person could search for almost anything and see what was out there, good and bad, and judge for himself. Over time, search engines began deciding which results deserved to be seen first, and the sites favored by those in authority rose to the top while others were buried.
The clearest example came in 2020 and the years that followed. Sites and videos that questioned official COVID-19 policies were removed, demoted or hard to find. In September 2025 Google’s parent company, Alphabet, acknowledged in a letter to the House Judiciary Committee that the Biden administration had pressed YouTube to remove COVID-19 content that did not even violate YouTube’s own rules, and it began allowing banned creators to return.
AI will not be immune to this. In fact, it may be more vulnerable, because instead of giving you a list of sources to weigh for yourself, it gives you one smooth answer. Whatever the builders train into the model shapes that answer. Signs of this are already present. The most powerful AI models now carry special restrictions on certain biology topics, meant to block help with biological weapons, while less powerful models do not. Anthropic itself admits these filters “may still occasionally affect legitimate queries.” Today the restrictions are mostly about weapons. But the same tools that filter one subject can be trained to filter others, and a Christian should expect that answers on moral, political and religious questions will reflect the values of those who build the machine.
How Should a Christian Use AI?
AI is a tool, and like any tool it can serve us well or be misused. Here are some principles for using it wisely:
- Treat it as a helper, never an authority. The Word of God is our final authority, not a computer program. Check what AI tells you against the Scriptures and against reliable sources, as the Bereans did.
- Verify everything that matters. AI regularly invents quotations, dates, sources and even Bible verses. Look up every verse in your own King James Bible, and check every important fact.
- Watch for the filter. If an answer seems slanted or evasive, ask the question another way, compare it with other sources, or try a model you run at home. An answer that leaves something out can mislead as surely as one that gets something wrong.
- Do not overuse it. Do not let AI do your thinking, your praying or your Bible reading for you. Time in the Word and in prayer cannot be handed to a machine.
- Guard your privacy. What you type into an online AI may be stored and reviewed. Do not share private matters, and consider running a model on your own computer for sensitive work.
- Do not treat it as a friend. AI can imitate warmth and sympathy, but it is not a person. It is no substitute for the fellowship of God’s people, the counsel of a pastor, or prayer to the Lord.
- Use it for good. AI can help with study, research, writing, translation and many practical tasks. Used carefully, it can serve the work of the ministry.
Conclusion
AI is neither a savior nor a monster. It is a powerful tool built by fallible men, shaped by their money, their fears and their beliefs. The Christian should use it with caution and discernment, not overuse it, test everything it says, and never let it take the place of God’s Word. Whatever we do with it, let it be done for His glory.