An urgent request can make checking feel like the wrong thing to do. Someone needs help. A delay might hurt them. A caller says there is no time to explain, or that telling another person will make things worse. The recipient is left to choose under conditions set by the person making the request.
A copied voice can add force to that pressure. Yet the pressure does not depend on a perfect copy. The claim of danger, the demand for secrecy and the borrowed authority of a relative or official can do work of their own. We need to understand that wider situation if we want a response that works when the audio is ordinary as well as when it is synthetic.
Consider an imagined call about a family member in trouble. The caller asks for an immediate payment and says nobody else must know. There are at least two tasks here: finding out whether someone needs help and deciding whether to send money to this caller’s chosen destination. The caller may present those tasks as one choice. We can separate them. Seeking help can begin while the payment remains unapproved.
That separation is the practical point. A person should not have to prove an entire story false before refusing to carry out an unverified demand.
What emotion research does and does not show
Kircanski and colleagues studied 231 adults in two age groups, 30–40 and 65–85. Participants received an emotion induction, then rated misleading advertisements and their willingness to buy. The results suggest that emotional arousal can affect susceptibility. But the stronger effects appeared among people who reached the intended emotional state; including those who did not weakened the results. The outcome was hypothetical purchase intention, not a telephone payment or measured financial loss. This study gives a reason to take the emotional setting seriously. It does not establish how much a pause or callback would help in a family emergency. Kircanski and colleagues, 2018
It would be a mistake to turn that result into a claim that emotion makes people foolish. Caring about another person is often the reason a request deserves attention. The design problem is how to give that care a safe route into action. Telling people to stop caring would fail both morally and practically.
There is a further problem with advice that says simply to remain calm. It puts the whole burden on the person who receives the frightening message. A usable agreement can do some of the work before the fear arrives. If unexpected payment requests always require a known-contact check, the recipient does not have to invent a new rule while upset. This is a proposed benefit of the agreement, not a measured effect from the emotion study.
Nor should age become a substitute for evidence. The study included younger and older adults. People of different ages also have different resources, responsibilities and needs. We learn little by calling one whole group gullible. We learn more by asking which task, pressure or missing support made a particular decision harder.
When a response is counted as success
A 2026 study by Heiding and colleagues used a matched sample of 4,100 people to examine responses to voice-phishing scenarios. Each participant encountered one recorded message or transcript, rather than a live back-and-forth call. In a cloned-sister condition with 107 participants, the reported weighted responses were 6.5% Yes and 29.6% Unsure. Adding them gives 36.1%. That combined figure is not the share who agreed to pay. The voice belonged to a researcher unfamiliar to participants, not to each listener’s sister. No real transfer was made. Heiding and colleagues, version 2, 2026
These differences matter. Being unsure, saying yes in a study, taking the first step in a transaction and losing money are different outcomes. A study can tell us something valuable about the first without measuring the last. Here it shows that some scenarios left people willing or uncertain. It cannot tell us the real-world loss rate, or isolate the effect of hearing one’s own relative.
It also leaves a question for prevention. What does an unsure person do next? They might pay. They might refuse. They might call someone who can establish the facts. A sound prevention study should observe that next choice instead of treating uncertainty itself as a completed failure.
This is why I place so much weight on giving people a route to check. The aim is not to make every person answer every message with instant certainty. The aim is to help someone act well while certainty is still missing.
Why ordinary trust is worth preserving
Timothy Levine’s truth-default theory offers a useful account of everyday communication. People often begin by accepting what they are told, unless something prompts doubt. This can make ordinary life efficient. We could not easily work, learn or care for others if each sentence required an investigation. The theory also distinguishes signs that raise suspicion from information that can settle whether a claim is true. It is a framework for understanding communication, not a tested family callback system. Levine, 2022
That distinction helps here. A demand for secrecy may give you a reason to check. It does not, by itself, prove fraud. A legitimate matter may be private. An urgent request may be real. The important step is to seek information that the original caller does not control.
We should also be careful with confident speech. A confident person may be wrong. Someone who sounds hesitant may be telling the truth. Making people perform confidence as a condition of being believed can punish exactly the person who is frightened or struggling to explain. Checking the claim through an established route is fairer than judging the speaker’s character from a pause.
Permission to question the request
The same issue appears at work. Imagine a staff member receiving what sounds like a senior manager’s instruction to change payment details. A written procedure will be weak if staff believe they will be punished for using it. A manager who wants checks to work must make room for the delay they require.
This is an organizational judgment, rather than a finding from the voice studies. The reason is straightforward: a rule that depends on a junior worker defying a powerful person will be fragile unless the organization backs the worker. A second approval should come through a route already established, with the particular destination checked. Asking the original caller to repeat the order supplies no new source of authority.
Within a family, the words can be less formal. An agreed response might be: I want to help. I am going to contact you in the way we agreed, and then we can work out what is needed. This is an illustrative script, not a statement that a particular sentence has been tested.
The script preserves both parts of the duty: care for the person and care with the decision. We should not make the recipient choose between them. If there may be immediate danger, seek help through the appropriate local emergency route. The payment demand can remain on hold while that happens.