Not legal advice. This site is an editorial reference. Laws change — always confirm with a qualified attorney in the relevant jurisdiction before recording, and check each page’s last reviewed date.

When to Record

The law tells you whether you may. Ethics is about whether you should. A short framework: necessity, consent, alternatives, and asymmetry.

The reflex to record has gotten cheaper. The phone is in your pocket; the storage is free; the audio quality is fine. The recording is no longer a thing you set up — it is a thing you tap. That change deserves more thought than it usually gets, because the consequences of a recording have not gotten cheaper. A conversation, once recorded, is an object: it can be edited, forwarded, leaked, subpoenaed, played back to people who were not in the room.

Necessity

The first question is whether the recording is necessary. Most calls do not need to be recorded. Most meetings do not need to be recorded. A good set of notes captures the substance with less of the cost. If you cannot say specifically why you need the recording — for a transcript you will actually read, for a difficult conversation you cannot remember accurately, for evidence you reasonably anticipate needing — the default should be not to record.

Consent

Even where the law permits one-party recording, the ethical default is the same as the legal default in strict-consent jurisdictions: ask. The question is two words: “mind if I record?” The answer is almost always yes. The ask costs nothing. The asymmetry between asking and not asking is real: a recording made with the other party’s knowledge is a different kind of object than a recording made without it.

There are exceptions. A reporter sourcing a confidential whistleblower may not have the luxury of asking. A person documenting harassment may not be able to ask. A person whose life is in danger may need a recording for a court they hope to bring later. These cases are real, and the legal landscape recognizes some of them through narrow statutory exceptions. But they are exceptions, not the default.

Alternatives

What does the recording let you do that a transcript, a contemporaneous note, or a follow-up email does not? A transcript of a phone call, written by you immediately after, captures most of the substance. An email confirming “just to recap our call: you said X, I said Y” serves as documentation that is harder to dispute and easier to share than an audio file.

The alternative is usually under-used. A recording feels like the safer move because it is a single artifact that contains everything. But the alternative — a written summary contemporaneously corroborated by the other party — is in many situations a stronger documentation that is also lighter on the other party’s privacy.

Asymmetry

Consider whether the people you are recording would record you if they were in your position. If the answer is yes — say, a sales call where you know the other side is recording you anyway — the asymmetry is low and recording is roughly neutral. If the answer is no — a personal conversation, a sensitive disclosure from a source, a moment of vulnerability — the asymmetry is high and the burden on the recording party is to justify it on something other than “the law allows me to.”

Asymmetry is not the same as power. A patient recording a doctor and a doctor recording a patient are differently situated; a journalist recording a public official and an official recording a private citizen are differently situated. The law mostly applies the same rule to both; the ethics does not.

The harder cases

Two cases recur. The first is the conversation you know is going to matter later: a salary negotiation, a difficult conversation with a relative, a confrontation with a contractor. Recording these is more often a substitute for the harder work of being a present, accurate listener than a complement to it. The second is the conversation you are about to publish: a podcast guest, an interview subject. The ethics here is largely solved by the release-form workflow, which makes both sides’ expectations explicit.

If the answer to whether to record is clear, record. If the answer is unclear, default to not recording — or to asking, which usually clarifies the answer.

A few concrete questions

  • What specific use will I make of the recording? Can I name it?
  • What is the cost to the other party if the recording exists, even if it is never shared?
  • Would I record this conversation if I had to tell the other party I was recording?
  • If the recording leaks, who is harmed?
  • Is there a less-recording version of this documentation that captures what I actually need?

Related