Bystanders pay a cost for being forced to overhear

What it is

A classroom is quiet. The phone beside you announces “dermatology at three, bring last week’s lab sheet.” The speaker pays for being heard. The bystander pays a different bill: attention seized, and knowledge they did not ask for. Eyes can turn from a screen; hearing has no equivalent “don’t look.” The bystander did not wake the device and did not consent, yet they pay the listening cost. This is not the speaker’s privacy loss. It is everyone else being pulled into a dialogue that is not theirs.

Why it happens

Auditory attention is a door that is hard to close in space: a mid-loud voice cuts reading and working memory even when the semantics are “not my business.” Dialogue output often carries full noun phrases and confirmation readbacks, denser than a muttered aside, so the intrusion is cleaner. Bystanders take two hits at once: a cognitive interrupt, and unwanted intimacy — a classmate’s skin, a colleague’s date, a child’s grade. They cannot angle a chair the way they dodge shoulder surfing; the options are headphones, a new seat, or endurance. Products that only ask owners “do you mind being overheard” delete this cost from the requirements.

Studying it

Bystander studies recruit the person in the next chair, not the person who woke the device. A confederate speaks to an assistant or lets it read aloud in a classroom or matched quiet place. Measures: interruption of the neighbor’s reading, annoyance, and ratings of “I now know something I should not.” Independent variables: output level, content class, whether the task needs quiet. Interviews ask whether bystanders want a veto (“kill that volume”), not only whether owners found it convenient. Owner satisfaction and bystander burden can worsen together — that split is the point.

Where it stops holding

Concerts, restaurants, already-loud streets: the increment of forced listening is small; speech is masked. Radios, roll-call, and ticket numbers define people present as legitimate listeners; bystander does not apply. People who chose the same meeting or the same car expect to hear others; the cost remains, the norm differs. Counting every overheard utterance as a violation would ban any sounding dialogue in public, including bus-stop announcements. The questions are whether the dialogue was addressed to the bystander, and whether they have a way off the channel.

Applying it

• In classrooms, open benches, waiting rooms — places with a quiet norm — default the speaker down or route to headphones. Living-room volume is not the default.
• Give people present an exit that does not depend on the owner: hardware mute, a visible “I am speaking” indicator, a one-tap kill on conference-room mics and loudspeakers.
• Treat bystanders as users in acceptance: in a quiet place, have an unaware neighbor read while a reply with calendar or health detail plays. Log interruption and the feeling of being stuffed with someone else’s private matter.
• If bystander annoyance is stable and the owner never notices, change level and content policy for the bystander, not for owner satisfaction.

Related

• Same group: M4.01.1 Speaking to a dialogue system hands the content to everyone in earshot · M4.01.2 Addressing a device in public is still socially marked · M4.01.3 Social cost drives people to abandon voice in public · M4.01.5 Spoken replies leak more than spoken commands · M4.01.6 Headphones privatize the reply, not the request
• Nearby: C7.07 Privacy visibility of voice input · M3.01 Naturalness and intelligibility of synthetic speech · M4.07 Felt privacy of always-on microphones
• Search terms: bystander forced listening · overhearing cost · bystander