For years, the public debate about artificial intelligence centered on what a machine could generate: words, images, computer code, recommendations, forecasts, or deceptive media. The next phase is different. The machine is being authorized to act. On September 17, the International Telecommunication Union published an article, prepared with World Bank Group participation, explaining how an AI agent could register a business, apply for a permit, seek government support, and arrange financing on behalf of a person. It identified the central question plainly: Is the agent genuinely authorized by the applicant? The article then stated that legal identity and digital ID now provide the foundation by which a person may delegate a task to an AI agent acting in that person’s name (Jamoussi & Vranic, 2026).
This is not a speculative technology demonstration hidden in a laboratory. The ITU’s Focus Group on Trust and Identity for Humans and Agentic AI is working on interoperable identity infrastructure for both people and autonomous agents. Its stated premise is that identity must now encompass dynamic, goal-directed software agents and must determine the conditions under which they can be trusted to act (International Telecommunication Union [ITU], 2026).
Commerce is moving in the same direction. Visa reported on September 9 that it had partnered with more than 100 companies to connect AI agents to its commerce network and that more than 150 card issuers were testing agent-initiated payments. Visa’s own survey found that only 23 percent of United States consumers trusted generative AI to handle payments, but 61 percent said they would trust the transaction when Visa secured it (Forestell, 2026). In July, Visa announced that AI agents had already completed live purchases with independent merchants in Europe through more than 30 participating card issuers. The system authenticated both the human cardholder and the AI agent acting for that person (Visa, 2026).
Mastercard has likewise introduced a system called Verifiable Intent. It links the consumer’s identity, the instructions given to an AI agent, and the agent’s resulting action in a tamper-resistant record. Mastercard describes the wider trajectory as one in which people delegate purchasing decisions to agents that interpret instructions, monitor conditions, and may act independently (Mastercard, 2026).
The important change is therefore not simply smarter AI. It is the construction of a trust architecture through which software can identify itself, prove who authorized it, receive permissions, cross institutional boundaries, and spend money. The Church should understand this development now, before convenience makes reflection seem unnecessary.
What My Earlier Work Established, and What Has Changed
My earlier research warned that digital identity was shifting from a means of confirming who a person is to a gateway determining what that person is permitted to do. In a comparative study of the European Union, India, China, and the United Kingdom, I argued that the convergence of digital identity, finance, and AI could make access to welfare, mobility, communication, and commerce increasingly conditional. I also insisted that privacy law, decentralized design, and viable alternatives for people who remain outside such systems are essential safeguards (Sangwa, 2025).
A related study examined programmable money, digital identity, standardized payment data, and AI-driven regulatory technology. That research found that these systems promise efficiency and inclusion while also moving gatekeeping power into the technical rails of economic life. It recommended offline or bearer-style payment options, explicit limits on programmability, data minimization, and governance arrangements that prevent any one institution from deciding who may participate in commerce (Sangwa & Mutabazi, 2025).
More recently, I described technopolar power as influence exercised through data, algorithms, cloud infrastructure, identity systems, payment rails, and online access, not only through territorial governments (Sangwa, 2026). I also examined the emerging programmable financial architecture and warned that a system can expand access while still creating new forms of dependence (Sangwa, 2026).
The newest evidence does not prove that those systems have become a unified instrument of global coercion. It does, however, materially advance the trajectory previously identified. The earlier concern involved the convergence of identity, money, and automated governance. The new development adds delegated machine action.
The emerging sequence is now clearer: human identity -> digital credential -> delegated mandate -> verified AI agent -> automated decision -> payment or public-service action -> machine-readable audit and compliance
This matters because the person is no longer merely presenting a credential to a system. A software representative may continually interpret the person’s goals, negotiate with other systems, and act within permissions assigned by governments, banks, platforms, and merchants. That is a genuine structural change.
What Is Verified, What Is Inferred, and What Remains Unknown
Prophetic watchfulness becomes unreliable when every inference is announced as a fact. We must therefore distinguish the levels of evidence.
Verified facts
AI agents have completed live purchases on behalf of cardholders in Europe.
Major payment networks are developing agent identities, tokenized payment credentials, delegated mandates, audit trails, and proof of user intent.
The ITU is developing international work on interoperable trust and identity for humans and agentic AI.
ITU and World Bank contributors are discussing AI agents that interact with business registries, licensing bodies, social-protection systems, lenders, and tax or health services.
The International Monetary Fund has identified agent identity, delegated authorization, programmable settlement controls, real-time compliance, and cross-border payment orchestration as parts of the emerging agentic-payment architecture (Davidovic & Tourpe, 2026).
Strong evidential inference
If adoption continues, AI agents will increasingly become intermediaries between people and institutions. Businesses are likely to move first because procurement, invoicing, reconciliation, travel, and supplier management are repetitive and rules-based. Consumer adoption will probably grow more slowly because trust remains limited. Public administration will also expand cautiously from low-risk functions such as form completion and status checks toward more consequential tasks.
This is an inference, but it is not guesswork. The payment networks, international standards bodies, governments, and financial institutions involved are openly building the necessary identity and authorization layers.
Plausible but unconfirmed concern
Voluntary agentic convenience may become de facto necessary if institutions redesign services around automated channels, if human service becomes slower or more expensive, or if fraud rules begin to distrust unverified users and agents. Interoperability can make services easier to use, but it can also spread one institution’s denial across many domains. A revoked credential, disputed identity, mistaken risk score, or politically expanded compliance rule could then affect several parts of life at once. The infrastructure makes this possible. Public evidence does not establish that governments and firms have already agreed to use it for universal ideological coercion.
What remains unknown
We do not yet know how widely consumers will adopt autonomous purchasing. We do not know whether competing identity systems will remain plural or converge into a small number of gateways. We do not know whether strong human alternatives will survive. We do not know how liability will be assigned when an agent misunderstands a command, favors the platform’s interest, or completes an irreversible transaction. We do not know how future governments will redefine prohibited or high-risk activity. These uncertainties are reasons for vigilance and wise policy, not permission to invent secret agreements for which we have no evidence.
The Biblical Foundation: A Human Being Is More Than a Credential
The first biblical question is not technological. It is anthropological: What is a human being? Genesis 1:26-27 teaches that God created humanity in His image. Human dignity is therefore neither granted by a state nor generated by a database. Government documents may identify a citizen for legitimate administrative purposes, but they do not create the person’s worth. A payment network may authenticate an account, but it does not define the soul. An algorithm may model preferences, but it does not know a person as the Lord does.
Psalm 139:1-16 describes God’s intimate knowledge of each person, including thoughts, words, and days. That divine knowledge is personal, righteous, and joined to the Creator’s authority. No institution should imitate omniscience while lacking God’s holiness, wisdom, and mercy.
For the believer, identity is also redemptive. Those who are in Christ are a new creation (2 Corinthians 5:17). They have received adoption through Christ (Romans 8:15-17), citizenship in heaven (Philippians 3:20), and a name known by the Lord (2 Timothy 2:19). The world may categorize believers by national number, credit profile, biometric template, risk score, purchasing history, or digital reputation. None of these can tell the deepest truth about them. This distinction is vital. A digital identity is a credential. It must never be confused with personhood. A software agent is an instrument. It must never be confused with a moral image-bearer.
Delegated Action Does Not Eliminate Human Accountability
Scripture recognizes delegated authority. Joseph administered Pharaoh’s resources (Genesis 41:39-44). Stewards managed households and property (Luke 12:42-48). Civil rulers appointed officers. The apostles entrusted responsibilities to qualified servants. Delegation itself is not sinful.
But Scripture never allows a person to outsource moral responsibility. Romans 14:12 says that each of us will give an account of himself to God. Second Corinthians 5:10 teaches that each person will appear before the judgment seat of Christ. A servant who carries out an unrighteous command does not make the master innocent, and a master who gives an unrighteous command cannot blame the servant. David remained accountable when Joab executed his orders concerning Uriah (2 Samuel 11:14-17; 12:9).
The same principle applies to agentic AI. If a Christian authorizes a system to deceive customers, exploit workers, manipulate prices, spread falsehood, purchase pornography, facilitate corruption, or discriminate unjustly, the Christian cannot answer God by saying, “The agent decided.” A machine cannot absorb our guilt.
This is one reason the current language of “delegated authority” deserves theological attention. An AI agent may operate at machine speed, but the human person, company, or institution that empowers it remains morally responsible. The command to do justice, love mercy, speak truth, use honest measures, and love one’s neighbor cannot be reduced to a technical permission file (Leviticus 19:35-36; Proverbs 11:1; Micah 6:8; Ephesians 4:25). Christians must therefore ask more than whether an agent completed a task efficiently. Was the goal righteous? Were the means truthful? Was the neighbor treated as a person rather than a data point? Who benefited? Who could appeal? Who bore the cost of the error?
Revelation 13: This Is Not the Mark, but It Expands an Enabling Condition
Revelation 13:11-18 joins several realities that interpreters must not separate: deceptive authority, false signs, an image associated with the beast, compelled worship, a mark connected to the beast’s name, and exclusion from buying and selling. The Mark of the Beast is not merely a digital identity, an AI agent, a credit-card token, a biometric login, or a payment standard. The text places the mark inside an idolatrous political-religious order. Receiving it expresses allegiance to the beast and rebellion against God (Revelation 14:9-12). To label today’s technology as the mark would go beyond Scripture.
Yet it would be equally irresponsible to ignore the mechanism described in the passage. The final system can condition economic participation on approved allegiance. It can identify those who comply and exclude those who refuse. Revelation does not reveal the engineering specification, but it clearly reveals the governing function: worship is enforced through access to commerce. Agentic identity strengthens a present enabling condition because it connects four forms of power:
Identification: Who is the human, organization, or software agent?
Authorization: What has that identity been permitted to do?
Execution: Can the agent act and transact without a fresh human decision at every step?
Compliance: Can the action be allowed, blocked, logged, reversed, or escalated according to machine-readable rules?
The IMF’s analysis is especially important here. It describes the movement from human “click-to-pay” transactions toward agent-mediated “decide-to-pay” processes. It also explains that rules concerning fraud, sanctions, identity, and policy can be enforced programmatically at the point of action. At the same time, it warns about opaque decisions, cybersecurity, ambiguous liability, concentration of power, and the danger that agents optimize provider incentives rather than user welfare (Davidovic & Tourpe, 2026).
The prophetic category that best fits the present evidence is therefore development of enabling conditions, not direct fulfillment. We are not witnessing Revelation 13 completed. We are witnessing an expanding capacity to bind identity, delegated action, compliance, and commerce together. A future beastly government would not need to invent every component from the beginning. It could appropriate systems already normalized for convenience, security, fraud prevention, and inclusion.
The Immediate Danger Is Formation, Not Only Future Coercion
Christians can become so focused on a future Antichrist that they miss the habits being formed now. Every technology teaches its users something. Agentic systems can teach people to prefer action without attention, convenience without understanding, and delegation without responsibility. They can encourage us to let machines choose what we read, buy, believe, prioritize, and ignore. The spiritual danger begins before any government mandate. It begins when human beings surrender judgment because discernment feels burdensome.
Romans 12:2 commands believers not to be conformed to this age but to be transformed by the renewing of the mind so that they may discern God’s will. Hebrews 5:14 describes mature believers as those whose powers of discernment have been trained through practice. That is the opposite of moral passivity.
An AI agent may compare prices faster than a person. It cannot decide what contentment requires. It may optimize a travel schedule. It cannot determine whether the journey honors God. It may recommend a charitable donation. It cannot love the poor. It may generate a prayer. It cannot approach the Father through the blood of Christ. It may summarize Scripture. It cannot obey Scripture, repent of sin, receive the Holy Spirit, or worship in spirit and truth. The more action we delegate, the more deliberately we must preserve human moral judgment under the Word of God.
Five Likely Developments the Church Should Watch
The following are responsible trajectory assessments, not claims of certain prophecy fulfillment.
1. From “know your customer” to “know your agent”
Financial systems already verify human customers. The next layer verifies the software acting for them. Agent directories, reputation systems, digital signatures, and revocable credentials will increasingly determine which agents may interact with merchants, banks, and public institutions.
2. From isolated tasks to continuous delegated authority
Early systems ask for confirmation at checkout. More advanced agents will operate under budgets, time limits, purchasing rules, and standing mandates. The critical question will be where bounded assistance ends and practical autonomy begins.
3. From private commerce to public administration
The ITU’s own examples include permits, social protection, health, tax, and business registration. The use of agents in government may improve accessibility and reduce paperwork. It may also make essential services dependent on compatible credentials, approved agents, and auditable digital interaction.
4. From national identity to cross-border recognition
The ITU is explicitly considering how an agent authorized in one agency or country may be recognized by another (Jamoussi & Vranic, 2026). Cross-border recognition can reduce friction, but it can also allow exclusions, risk designations, or technical failures to travel farther.
5. From optional convenience to infrastructural dependence
No one needs to announce a universal mandate for dependence to grow. If the cheapest prices, fastest permits, best customer service, most efficient business tools, and preferred payment channels all assume agentic participation, the person who refuses may remain legally free while becoming practically excluded. This is how infrastructures acquire power: not only by issuing commands, but by becoming difficult to live without.
A Christian Standard for Evaluating Agentic Systems
The Church should neither reject every advanced technology nor receive it uncritically. The biblical test should include at least these questions:
Human dignity: Does the system serve image-bearers, or reduce them to profiles that may be scored and excluded?
Truth: Can a user understand what the agent did, why it did it, and whose interests it optimized?
Accountability: Is there a responsible human or institution who can answer for harm?
Consent: Is authorization specific, informed, revocable, and limited in duration and scope?
Justice: Can the poor, elderly, disabled, undocumented, falsely flagged, or technologically excluded still obtain essential services?
Mercy: Is there a meaningful human appeal when automated enforcement is wrong?
Privacy: Is only the minimum necessary information disclosed?
Freedom of conscience: Can a person decline the agentic channel without losing the ability to buy, work, travel, communicate, worship, or receive care?
Plurality and resilience: Are there offline, cash-like, local, and non-platform alternatives?
Spiritual authority: Is the tool performing a bounded task, or becoming a guide that shapes faith, conscience, identity, and obedience?
These are not anti-technology questions. They are pro-truth, pro-neighbor, and pro-human questions.
What Churches, Families, and Christian Institutions Should Do Now
First, pastors should teach a biblical doctrine of humanity. Christians who do not know that identity is grounded in creation and redemption will easily accept identity as something conferred by platforms, institutions, or synthetic personalities.
Second, churches should teach moral responsibility in an automated age. Delegation does not cancel stewardship. Believers should be trained to review what their systems do, restrict permissions, retain human approval for consequential actions, and refuse uses that violate Scripture.
Third, Christian institutions should practice data restraint. A church does not need to collect every available data point merely because software makes collection easy. Sensitive pastoral, financial, family, and discipleship information should not be poured casually into agentic platforms.
Fourth, Christian leaders should advocate for human alternatives. Governments and companies should preserve cash and offline options, accessible human service, narrow purpose limitations, revocable consent, transparent appeals, and protection against discrimination based on lawful beliefs.
Fifth, households should develop practical resilience. This does not mean panic or withdrawal from society. It means reducing unnecessary debt, keeping essential records, maintaining direct relationships, learning useful skills, practicing generosity, and strengthening church-based mutual aid. A congregation that only meets through platforms it does not control is more vulnerable than one joined by actual covenantal love.
Sixth, parents should teach children that convenience is not the highest good. Young people need practice making decisions, evaluating claims, managing money, reading Scripture in context, speaking to real people, and accepting responsibility for choices. A child trained to outsource every difficult task may become an adult unable to resist persuasive automation.
Seventh, believers working in technology, banking, government, and standards bodies should remain present and faithful. They can press for limited permissions, privacy-preserving design, human appeal, transparent audit, decentralized safeguards, and non-discriminatory access. Joseph and Daniel served inside powerful administrations without surrendering worship. When ordered to cross God’s boundary, however, Daniel refused (Daniel 6), and the apostles declared, “We must obey God rather than men” (Acts 5:29).
The Gospel Question No System Can Answer for You
The greatest danger is not that an AI agent may spend your money incorrectly. The greatest danger is that you may stand before God unforgiven. No digital identity can register a person in the Book of Life. No biometric credential can prove righteousness before God. No agent can repent in your place. No algorithm can bear your sin. No compliance score can reconcile you to your Creator.
Scripture teaches that all have sinned and fall short of the glory of God (Romans 3:23). The wages of sin is death, but the gift of God is eternal life in Christ Jesus our Lord (Romans 6:23). Jesus Christ, the eternal Son of God, died for our sins and rose bodily from the dead (1 Corinthians 15:3-4). Salvation is received by grace through faith, not earned by works, rituals, technology, religious affiliation, or moral scoring (Ephesians 2:8-9).
The urgent call is therefore to repent and believe the gospel. Those who belong to Christ are forgiven, adopted, sealed by the Holy Spirit, and kept for an inheritance that cannot be corrupted or revoked by earthly administrators (Ephesians 1:13-14; 1 Peter 1:3-5). The pre-tribulational blessed hope teaches the Church to wait for Christ Himself, not for the completion of a technological checklist. We do not know the day or hour of the rapture, and we must not set dates. We do know that our Lord calls us to remain awake, holy, faithful, and ready (John 14:1-3; 1 Thessalonians 4:13-18; Titus 2:11-14).
Conclusion: Do Not Surrender the Name by Which You Must Answer to God
The most timely warning is not that AI agents are secretly the Mark of the Beast. They are not. The warning is that identity, authority, action, compliance, and commerce are being joined in a new technical layer. The latest ITU and payment-network developments show that this layer is moving from theory into standards, live transactions, and public-service planning. It may bring genuine benefits. It may also create unprecedented capacity for invisible gatekeeping, automated enforcement, and cross-domain exclusion.
The Church must not answer with panic. We must answer with truth. Use tools without worshiping them. Delegate tasks without surrendering conscience. Demand accountability without pretending technology can be morally neutral. Preserve alternatives for those who cannot or will not enter a system. Defend the poor from automated exclusion. Refuse every command that requires disobedience to Christ. Test every promise of safety, inclusion, and efficiency by the Word of God.
Above all, remember whose name truly matters. The world may identify you by numbers, credentials, devices, agents, and accounts. Jesus said that His sheep hear His voice, He knows them, and they follow Him (John 10:27-29). Be certain that you belong to Christ. The decisive question is not whether an AI agent can act in your name. It is whether your name is written in heaven (Luke 10:20; Revelation 20:15).
Recommended Reading
Sangwa, S. (2025). Global Digital Identity Systems: Interdisciplinary Insights into Surveillance, Governance, and Prophetic Implications.
Sangwa, S., & Mutabazi, P. (2025). Programmable CBDCs, Digital Identity, ISO 20022, and AI: Reconfiguring Power in Payment Infrastructures.
Sangwa, S. (2026). The Money of Tomorrow Is Being Designed Today: Jackson Hole, Elite Policy Networks, and the Biblical Test of Financial Power.
Sangwa, S. (2026). After the Digital Euro Vote, Are We Seeing Digital Cash or Credentialed Participation?
Sangwa, S. (2026). Could Modern Technopolarity Be Preparing the World for the Final Beast System?