One faith, several builders
The Seventh-day Adventist Church is developing more than a single chatbot. Regional church departments, the General Conference and independent Adventist developers are building different ways to search religious documents, explain beliefs and prepare ministry resources. Understanding who owns each tool is becoming as important as understanding what it can answer.
The General Conference’s October 13 presentation of AKAI and Adventist+ supplies the immediate news context. Its October 15 announcement identifies AKAI as a church-owned information platform developed with Hope Channel International, while explicitly withholding the status of official church voice or spiritual authority. It describes a family of tools, not a replacement for every existing project. Source: adventist.news
Our earlier report examined the announcement’s human prayer referrals. This analysis addresses a different question: who builds these systems, which sources they privilege, and what users should understand about institutional endorsement. Source: aifaithmonitor.org
A denominational name on an interface can communicate familiarity. It cannot, by itself, tell a reader whether the software belongs to the worldwide church, a regional administration, a supporting organization or a private developer.
Regional projects came before the latest announcement
The South American Division launched 7chat.ai in beta on May 1, 2025. Its Digital Strategies department described an assistant for Bible questions, church manuals, studies, sermon ideas and other ministry material. The announcement acknowledged possible inaccuracies and advised users to check official sources and consult church leaders about doctrinal doubts.
The same report traces an earlier initiative, Esperanza, to the Nuevo Tiempo communication network in 2017. Initially a Bible-study chatbot, it later gained generative capabilities for open-ended questions using Adventist archives. The dates should remain separate: an older assistant’s original launch does not establish when its current AI features became available. Source: adventist.news
In Korea, Adventist Church GPT was publicly reported in June 2024. The Korean Union Conference’s Digital Evangelism department presented it as a way to explain doctrine, history and lifestyle and support Bible-school interactions. Its launch coverage describes training on church documents and technical support from OPENBUS. Favorable beta-user comments appear in that coverage, but selected testimonials are not an independent accuracy study. Source: adventist.news
These histories establish multiple institutional initiatives. They do not establish that all three share one model, privacy policy or system of oversight.
AdventAI and Adventist Intelligence require separate labels
AdventAI describes itself as a search and study tool for the Bible, Ellen G. White’s writings and early Adventist literature. Its team page identifies Emmanuel Higgins as founder, Wyatt Allen as co-founder and Eric Louw as founding partner, and associates the project with Blue Vineyard Group. That documentation describes a developer-led project; it does not establish General Conference ownership. Source: adventai.app
A separately branded service, Adventist Intelligence, states that it is built in coordination with ASI North America. Its FAQ explicitly says it is not an official Seventh-day Adventist Church product. The site describes restricted access before a wider public opening, linked sources and a mechanism for reporting errors to a human reviewer. These are the operator’s descriptions, not findings from our own performance tests. Source: adventistintelligence.com
The distinction is easy to lose because the words Adventist, intelligence and AI recur across projects. Readers should not infer organizational identity from similar names or treat a supporting association’s involvement as worldwide denominational endorsement.
A useful directory therefore needs separate fields for operator, institutional relationship, intended audience, source collection and current availability. Calling everything “the Adventist AI” conceals information readers need.
The stated motive is more accurate religious representation
Why create specialized systems when general chatbots already exist? In Amanda Blake’s May 2026 reporting for Columbia Union Visitor, Adventist Learning Community director Adam Fenner points to inaccurate general-AI answers about the denomination. Richard Stephenson similarly describes a desire to make verified Adventist material accessible. Those are attributed reasons for development, not proof that specialized products have achieved superior accuracy. Source: columbiaunionvisitor.com
The distinction between representing a tradition and establishing religious truth is essential. A tool may accurately explain what a denomination teaches without resolving every theological disagreement about that teaching. Readers should know which task the assistant is attempting.
For example, someone might ask what Adventists teach about the Sabbath. A denominational assistant can help locate the relevant statement. A comparative question about why Christians disagree requires recognition of other traditions and their own sources. A narrow library may be useful for the first task while insufficient for the second.
This is not an argument against specialized religious tools. It is an argument for describing their scope honestly, so users can distinguish a well-supported institutional explanation from a comprehensive account of a contested subject.
Who chooses the library shapes the answer
The Adventist Technology Summit’s published program describes AKAI’s use of retrieval-augmented generation, source curation and governance. Its session outline treats the selection of documents and the enforcement of disclosure rules as architectural decisions. The program is evidence of the project’s described approach, not a published independent evaluation. Source: ats.adventist.org
In ordinary terms, retrieval brings relevant material into the answer-generation process. The reader should still ask whether the cited passage supports the resulting sentence. Selecting a reputable source does not settle whether a summary preserves its meaning.
Consider a hypothetical answer combining a historical encyclopedia entry, a current doctrinal statement and an individual sermon. Those documents serve different purposes. A historical account describes developments; an adopted statement defines an institutional position; a sermon contains a speaker’s interpretation. A responsible interface should not make them appear interchangeable.
The public Adventist Intelligence site is unusually explicit about a hierarchy: Scripture, church teaching, Ellen White and other Adventist voices. This is a declared editorial ordering within that particular product, not evidence that all Adventist systems use the same hierarchy. Source: adventistintelligence.com
Readers need more than a reassuring reference number. They need the original passage, enough surrounding context and a clear distinction between quotation and generated explanation.
What independent research can and cannot tell us
A September 12, 2026 paper by Marcos Salvatierra in the International Journal of Latin American Religions examines 7chat.ai through public institutional documents. It explores how the language and design surrounding denominational AI could affect religious authority, expectations of pastoral care and tolerance for ambiguity.
Its limits are important: this is document-based analysis, not a trial measuring actual users’ behavior. It does not establish that the tool has caused manipulation, dependency or other harms. Its arguments identify questions for empirical investigation. Source: link.springer.com
Independent journalism adds a concrete implementation issue. Andrew Francis’s October 15 Spectrum report says human staffing was envisaged for OneVoice27, but a continuation plan after the initiative’s September 2027 conclusion had not yet been developed. That is the reported position at the presentation, not a claim that the question will remain unresolved. Source: spectrummagazine.org
Together, these sources encourage scrutiny without justifying a verdict against the technology. We need evidence about actual use, correction and oversight, rather than treating either promotional assurances or theoretical risks as established outcomes.
Governance is part of the religious service
The South American Division published ethical principles on November 7, 2023, addressing transparency, fairness, reliability, accountability, privacy, security and sustainability. The document treats AI as auxiliary to human relationships and anticipates more specific guidance for different areas of church activity. It is a regional framework, not proof that every Adventist product follows one independently verified global policy. Source: institucional.adventistas.org
For a congregation choosing a tool, that framework suggests practical questions. Who maintains its sources? How are outdated documents identified? Who reviews a disputed answer? Can users understand what happens to their conversations before sharing personal concerns?
These questions matter even when a service is free. A query about religious history may disclose little about the user. A conversation about family conflict or loss of faith can reveal far more. A church should explain the difference between searching a library and entrusting private information to a service.
The answers should be product-specific. Neither shared religious identity nor a familiar institutional name removes the need to inspect ownership, policies and the route to human accountability.
Christian perspective: a source is an invitation to examine
Christian discernment offers a constructive standard for these tools. Acts 17:11 describes the Bereans examining Scripture to assess the message they received. Applied to religious AI, that example favors systems that make checking easier rather than encouraging deference to fluent answers. Acts 17:11 (NIV)
A helpful assistant might direct a learner to an unfamiliar passage or recover a document that was difficult to find. The next step should deepen understanding: read the chapter, consider its setting and ask whether the explanation fits. Efficiency serves learning when it opens that process; it weakens learning when it replaces it.
James 3:1 emphasizes the responsibility borne by teachers. In this setting, that responsibility belongs to the people who choose, deploy and distribute a tool’s answers. Software cannot absorb the moral accountability of a pastor or institution that presents material to a congregation. James 3:1 (NIV)
This gives the growing family of Adventist AI projects a demanding but useful measure of success. More answers are not enough. The tools should help people identify their sources, understand the limits of institutional claims and remain willing to examine difficult questions with accountable human teachers. A recognizable religious name should begin that inquiry, not end it.