A significant transition is taking place in the digital world. Governments and financial institutions are no longer developing identity systems only for human beings. They are beginning to build identities, authorization rules, trust registries, and payment credentials for artificial-intelligence agents that can act on behalf of people and organizations.
This does not mean that artificial intelligence has become human. It does not mean that AI possesses a soul, bears the image of God, or deserves spiritual trust. It does not establish that the mark of the beast has arrived.
It does mean that digital systems are acquiring a new capability: a machine can be recognized, authorized, audited, and permitted to transact within an expanding identity-and-payment infrastructure.
That development deserves careful Christian attention. In earlier articles, I argued that digital public infrastructure is becoming a reusable participation layer joining identity, payments, data exchange, public services, and programmable rules. I also examined how technology companies increasingly exercise powers once associated mainly with governments (Sangwa, 2026a, 2026b).
The latest evidence permits a more precise conclusion. The emerging architecture is not merely learning to identify a person before allowing a transaction. It is learning to recognize a digital agent, establish whose authority it carries, determine what it may do, and allow it to spend money within machine-readable limits.
Estonia’s Proposed Identity Codes for AI Agents
In June 2026, Estonia’s government announced plans to develop official digital identities for AI agents. These “AI ID codes” are intended to let an agent act for a person, company, or organization while remaining separately identifiable and subject to defined permissions.
The Estonian proposal envisions distinguishing whether an agent may view information, prepare a document, initiate a payment, or operate only below a specified financial limit. Its stated purpose is accountability: systems should be able to determine which agent acted, for whom it acted, what authority it possessed, and who remains responsible for the result (Government Office of Estonia, 2026).
There are legitimate reasons for such controls. Giving an AI assistant unrestricted access to a person’s accounts, files, communications, and financial credentials would be extremely dangerous. A separate identity with revocable, limited permissions may be safer than allowing an agent to impersonate its human operator.
Nevertheless, the deeper significance should not be missed. Estonia is proposing a model in which artificial agents become recognized participants within a national digital trust environment. They do not become citizens or persons, but they acquire something resembling an administrative passport: a credential that allows a system to recognize their delegated authority.
From Experimental Agents to Authorized Economic Actors
This development is not occurring in isolation. The U.S. National Institute of Standards and Technology has been examining how identity, authentication, authorization, delegation, and audit requirements should apply to AI agents. Its 2026 concept paper specifically considers how systems can distinguish human identities from agent identities, bind an agent to a human authorization, restrict the agent’s entitlements, and maintain verifiable records of what it does (National Institute of Standards and Technology, 2026).
The OpenID Foundation has likewise advanced authorization frameworks for situations in which an agent cannot act until a prerequisite has been satisfied, such as human approval, consent, delegated authority, an attestation, or a risk assessment (OpenID Foundation, 2026).
Meanwhile, payment networks are moving agentic commerce from theory toward live operation. Mastercard reported that European banks had completed controlled live transactions in which AI agents acted with consumer authorization. Its framework uses tokenization, authentication, traceability, and verifiable intent to connect an agent’s action to the consumer who approved it (Mastercard, 2026).
Visa has described agents that book travel, reorder inventory, purchase computing services, and operate within assigned budgets. It is also developing an agent directory, behavioral trust signals, tokenized credentials, and infrastructure for AI-initiated payments. Visa’s announced direction is to place identity, permissions, and behavioral signals more deeply inside payment credentials (Visa, 2026a, 2026b).
These corporate statements naturally promote the companies’ own products. They should not be mistaken for independent proof that every promised safeguard works. They do, however, provide primary evidence of the direction in which major payment institutions intend to move.
Human Wallets and Machine Agents Are Converging
At the same time, digital wallets are becoming more closely connected to government credentials, private services, payments, and cross-border trust frameworks.
The World Bank’s 2026 digital-wallet policy work describes the convergence of digital identity, verifiable credentials, data sharing, electronic signatures, and payments within wallet ecosystems (World Bank, 2026a). Its related work on instant payments explores connecting authoritative digital identity credentials to onboarding, transaction authorization, fraud prevention, and payment-history sharing (World Bank, 2026b).
The World Bank says its broader Digital Public Infrastructure program supports digital identity, payments, and secure data sharing in more than 80 countries. It promotes these components as interoperable foundations for social protection, finance, health, agriculture, employment, and public administration (World Bank, 2026c).
In Europe, the eIDAS Dashboard is developing into part of the trust framework supporting the European Digital Identity Wallet. The infrastructure includes machine-readable lists of recognized wallet providers, identity providers, attribute providers, age-verification providers, and other trusted entities. The European Commission also reports work toward compatibility with countries outside the European Union (European Commission, 2026).
In the United Kingdom, government guidance updated on September 1 points identity providers toward a GOV.UK Wallet sandbox. The wallet presently holds a veterans credential, with a digital driving licence intended to follow. Certified services will be able to use government-issued wallet information to create derived credentials or prove attributes such as eligibility and age (Government Digital Service & Office for Digital Identities and Attributes, 2026).
Individually, these developments have practical and often beneficial purposes. Together, however, they reveal an emerging structure: Human identity → digital credentials → delegated AI authority → automated decisions → programmable payment → auditable participation.
No document examined establishes a single secret command directing every initiative. The evidence instead shows technological convergence, institutional diffusion, shared standards, and overlapping commercial incentives.
That distinction matters. Christians must not invent coordination where documentation establishes only convergence. But neither should we ignore convergence merely because it is not controlled from one room.
A Digital Identity Is Not the Image of God
Scripture begins the discussion of identity somewhere entirely different from modern governments and technology companies. Human beings are made in the image and likeness of God (Genesis 1:26–27). God formed man from the dust and breathed into him the breath of life (Genesis 2:7). Human dignity does not arise from a credential, database entry, biometric match, citizenship status, economic score, or digital reputation.
An AI agent may receive a technical identity, but it does not thereby receive personhood. It may generate persuasive language without possessing wisdom. It may simulate concern without loving its neighbor. It may execute an instruction without bearing moral responsibility before God.
Scripture gives no basis for treating a human artifact as a spiritual person. Romans 14:12 says that each of us will give an account of himself to God. A person cannot transfer that final accountability to an algorithm.
This has immediate implications. Christians must not allow machines to become substitutes for conscience. An AI system may help analyze information, but it cannot repent. It may suggest a donation, but it cannot love the poor. It may prepare a sermon, but it cannot receive the illumination of the Holy Spirit. It may imitate pastoral language, but it cannot shepherd a soul as one who must give an account to God (Hebrews 13:17). The danger is not that the machine secretly becomes human. The danger is that humans surrender increasingly human responsibilities to machines.
Revelation 13: Capability Is Not Fulfillment
Revelation 13 describes a final system joining political authority, spiritual deception, idolatrous worship, an image that speaks, a compulsory mark, and economic exclusion. No one may buy or sell without the mark connected to the beast’s name or number (Revelation 13:15–18).
The controlling issue is worship and allegiance. Therefore:
An AI identity code is not the mark of the beast.
An AI payment is not the mark of the beast.
A government wallet is not automatically the mark.
A digital credential is not, by itself, worship of the Antichrist.
Estonia is not proven to be constructing the beast system.
Visa, Mastercard, the World Bank, the European Commission, and NIST are not thereby proven to be knowingly coordinating an Antichrist program.
Those claims would go beyond the evidence and beyond Scripture. Yet the same biblical discipline forbids naivety. Revelation foretells real economic exclusion. A future authority must be able to distinguish permitted participants from prohibited participants and enforce that distinction across ordinary commerce.
Identity-linked wallets, agent authorization, interoperable trust registries, programmable payments, behavioral risk systems, and automated compliance do not fulfill that prophecy. They create capabilities compatible with the kind of economic permissioning Revelation describes. The proper category is therefore development of enabling conditions, not direct fulfillment.
The New Chokepoint: Who Defines the Agent’s Mandate?
The most important question is not whether an AI agent has an identification number. It is who defines what that identity is permitted to do. An agent may be programmed not to purchase an unlawful product. That may protect society. But what happens when authorities redefine biblical teaching as harmful extremism? What happens when a Christian book, ministry, school, medical decision, charitable transfer, or missionary payment is classified as unsafe?
The same architecture that prevents fraud could restrict lawful conscience. The same trust registry that excludes malicious agents could exclude disfavored organizations. The same behavioral signal that catches stolen credentials could become a compliance score. The same automated assistant that finds the cheapest product could quietly refuse transactions that conflict with its platform’s moral policies.
A further trend is likely: people may increasingly be expected to interact with institutions through authorized agents. Banks, governments, insurers, employers, schools, and marketplaces may prefer machine-readable requests because machines can process them quickly. Those without an accepted identity, compatible wallet, approved agent, or sufficiently trusted credentials could then experience practical exclusion even without an explicit law banning them. Economic control does not always begin with a public prohibition. It can emerge through accumulated dependencies and default settings.
Daniel’s Warning About Delegated Obedience
Daniel 3 presents a political image, a public command, music that signals the required response, and a penalty for noncompliance. The technological details differ completely from our world, but the spiritual pattern remains instructive: administrative coordination becomes beastly when it demands what belongs to God.
Shadrach, Meshach, and Abednego did not wait until the furnace was visible to form their convictions. Their answer had already been settled: God was able to deliver them, but even if He did not, they would not worship the image (Daniel 3:16–18).
Daniel 6 makes the same point. Daniel’s prayer life was established before prayer became administratively punishable. When the law changed, his loyalty did not.
Christians should therefore ask now: What decisions will we never delegate to a machine? What information should our churches refuse to centralize unnecessarily? Which payment, communication, and hosting alternatives should ministries preserve? Are our members being discipled to obey God when convenience and faithfulness separate?
What the Church Should Do Now
First, teach biblical personhood. Children and adults must understand why a fluent machine is not a soul, prophet, pastor, or spiritual authority. Second, preserve meaningful human approval. Churches, ministries, schools, and Christian businesses should require accountable human decisions for pastoral care, discipline, doctrinal publication, significant financial transfers, employment decisions, benevolence, and other matters affecting dignity or conscience.
Third, apply minimum permissions. An AI tool should receive only the data and authority required for a defined task. Access should be revocable, financial limits should be explicit, and important actions should be reviewable.
Fourth, maintain lawful alternatives. Ministries should avoid total dependency on one identity provider, cloud platform, payment processor, AI system, or communication channel. Practical resilience may include multiple payment methods, careful cash policies, local mutual aid, offline contact records, and contingency planning.
Fifth, defend people who cannot participate digitally. James condemns partiality toward the socially powerful (James 2:1–9). A church must not treat the elderly, poor, disabled, undocumented, technologically inexperienced, or conscientious objector as administratively inconvenient.
Sixth, refuse false witness. Do not turn every standards meeting into a satanic council or every wallet into the mark. Exaggeration weakens the watchman’s warning. “The simple believes every word, but the prudent considers well his steps” (Proverbs 14:15).
Seventh, prepare spiritually. The greatest threat is not that an AI agent might spend our money. It is that convenience, fear, prosperity, social acceptance, or economic pressure may reveal that Christ never possessed our hearts.
Our Final Identity Is in Christ
A person can possess every approved credential and remain lost. A person can understand digital surveillance, reject centralized systems, and still die without Christ. The gospel is not an escape from technology. It is the good news that Jesus Christ died for our sins, was buried, and rose again according to the Scriptures (1 Corinthians 15:1–4). Salvation is by grace through faith, not through technical knowledge, political resistance, or prophetic research (Ephesians 2:8–10).
Repent and believe in the Lord Jesus Christ. Be reconciled to God while the invitation remains open. Believers should pursue holiness, proclaim the gospel, test every spirit, and wait expectantly for Christ, who will gather His Church before the coming Tribulation (1 Thessalonians 4:13–18; Titus 2:11–14).
AI agents are receiving identifiers. Digital wallets are becoming participation gateways. Payment systems are becoming increasingly programmable. These facts do not prove that Revelation 13 has been fulfilled.
They do remind us that the economic mechanisms described by John are no longer administratively unimaginable. The Church must neither panic nor sleep. We must understand the architecture, tell the truth about the evidence, refuse premature prophetic identifications, preserve human responsibility, prepare for conscience pressure, and remain faithful to Jesus Christ. The decisive question is not whether your agent will be authorized. It is whether you are known by Christ.
Recommended Readings
Sangwa, S. (2026). Could Digital Public Infrastructure Become the Backbone of Conditional Participation?
Sangwa, S. (2026). Could Modern Technopolarity Be Preparing the World for the Final Beast System?
Sangwa, S. (2025). FAQs: What Is the Dark Agenda Behind Digital ID Systems?
Sangwa, S. (2025). Will Most People Accept the Mark of the Beast?
Sangwa, S. (2023). What Is the Mark of the Beast and How Do I Know If I Have It or Not?