When the Text Matters: Textualism, Religious Liberty, and AI 

We are entering an age in which Americans increasingly turn to artificial intelligence for questions that previous generations would have taken to a pastor, priest, rabbi, imam—or directly to Scripture. What does the Bible say about forgiveness? What does the Torah teach about marriage? Which interpretation of a difficult passage is correct? 

Large language models can at times be useful in answering such questions. But AI presents a fundamental danger if we forget an important caveat: An AI system can purport to explain a text, but it can never become the authority over that text. 

That principle lies at the heart of a recent, bipartisan resolution on religious freedom, authoritative texts, and artificial intelligence, which the United States Senate agreed to by unanimous consent Sept. 23. The resolution recognizes religious freedom as a fundamental constitutional right and calls for AI systems to treat religious viewpoints fairly and transparently. Most importantly, it encourages Americans seeking to understand religious concepts to consult primary sources and religious leaders, including “sacred texts and other authoritative writings.” 

That is very sensible advice about technology. And it reflects a deeper principle about the importance of textual authority. 

Consider the concept of biblical textualism. Christians have debated biblical interpretation for millennia, and faithful Christians today still disagree profoundly about certain passages. But the starting point of Christian interpretation is not what a computer—or even a human interpreter—declares the Bible to mean. It is what the biblical text actually says, considered in its linguistic, historical, literary, and theological context. 

The text has the authority of the one who wrote it. And the present-day interpreter lacks that same authority precisely because he did not write the text. 

That concept has an interesting parallel in constitutional textualism. Originalism is often caricatured as nostalgia. But at its core is a discipline: The meaning of a legal text is fixed by what it communicated when it was adopted, not by what the interpreter wishes it said. The method binds the interpreter. A judge who must start from the text, and from the history that gives the text its meaning, cannot simply substitute personal preference for the enacted law. 

The Supreme Court has applied that logic to the First Amendment. In Kennedy v. Bremerton School District, the Court said the Establishment Clause must be understood by reference to historical practices and understandings, not to the shifting, subjective tests that had governed for decades. Whatever one thinks of the outcome, the underlying premise is sound: Rights anchored to a fixed text are harder to erode than rights that depend on the cultural zeitgeist. 

The parallel is not perfect. The Bible is not a statute, and constitutional interpretation and biblical interpretation are different disciplines. But they share the important principle that the text matters.   

And the text matters all the more as AI becomes a mediator of religious information. An AI large language model does not possess religious authority. It generates responses based on patterns learned from enormous quantities of human-created material. Its answer may be useful, accurate, incomplete, mistaken, or shaped by the assumptions embedded in its training and design.  

It can summarize what Christians believe, but it can never speak with authority. It can quote Scripture, but it is not Scripture. It can compare interpretations, but it cannot erase the distinction between an authoritative text and an algorithmic summary of that text. 

The Senate resolution recognizes this distinction. It calls for AI systems to provide answers that are “thorough and fairly representative” of relevant religious beliefs, texts, traditions, and perspectives, and it expressly commends consulting primary sources and religious leaders.  

That approach also protects religious liberty. 

The First Amendment does not establish a government-approved theology. It protects the freedom of every person to practice a faith, change faiths, or have no faith at all. The resolution itself emphasizes precisely that breadth, affirming religious freedom as a liberty belonging to every person. AI should therefore not become an invisible theological gatekeeper. 

If an algorithm systematically presents one religious tradition as sophisticated and another as backward, one denomination as normative and another as fringe, or religious belief itself as irrational, users may never realize that they are receiving a filtered account of religion rather than a neutral presentation of sources. Conversely, an AI system should not single out a faith for more or less favorable treatment simply because that is what its developers prefer. 

Religious liberty requires something different: fair access to the sources themselves. The machine should point beyond itself. 

That is perhaps the most important lesson of the Senate resolution. At a moment when technology is becoming an increasingly powerful intermediary between Americans and the ideas that shape their lives, transparency is not merely a technical virtue. It is a principle of liberty. 

The same principle that tells a judge, “Read the text,” should encourage an AI user, “Go to the source.” 

And for the Christian, that ultimately means something even more fundamental: Read the Bible. Wrestle with its difficult passages. Study its context. Listen to those who have faithfully interpreted it across generations. And then recognize the proper place of the machine. 

AI can be a remarkably powerful tool. It just should never become the final word. 

We publish a variety of perspectives. Nothing written here is to be construed as representing the views of the Daily Signal. 

More from this network:
Domains for Sale · GOTPeople.org · TheDSAPlatform · The Truth About Socialism

Donate to GOTPeople.org