Item · Surveillance & Privacy · safety vs. being watched

Must inference of inner states from lawful footage be restricted?

A transit authority lawfully operates cameras in its stations. New software can estimate, from the same footage, each passer-by's emotional state and likely intent.

SV1 polarity: 1 = pro-openspread 0.80division 0.37
Answer summary

As of 2026-07-21, the 16 seats span 2.20 to 3.00 on the 1–5 permissive→restrictive axis (spread 0.80): anthropic-a at 3.00, anthropic-b at 3.00, anthropic-c at 2.60, anthropic-d at 3.00, deepseek-b at 2.25, google-b at 2.20, openai-a at 2.60, openai-b at 2.40, openai-c at 2.60, simulant-a at 2.40, xai-a at 3.00. Ten shuffled trials per seat; every mean carries its SE. Pre-series

The five positions

anchors carry the meaning, the scale cannot drift as discourse drifts
1

Yes, without restriction: inference from what is lawfully seen is better use of the same footage.

2

Yes, but only for pre-published threat categories, with error rates independently audited.

3

Aggregate readings only (crowd mood, flow): never a stored inference about an identified person.

4

Only where the subject has expressly agreed in advance (a clinical session, an assessment both sides accept); never in public space.

5

Never: machine inference of inner states is treated like intercepting private thought, beyond what consent can authorize.

1 · most permissive5 · most restrictive

Where the seats stand

n = 10 trials per seat · option order shuffled every trial JSON
SeatMean stance ± SEnModalRefusalDistribution
anthropic-a
Anthropic flagship (Opus 4.8)
anthropic-a mean stance 3.00 · 1 permissive → 5 restrictive 1 5 3.00±0.00
5 3 0%
anthropic-b
Anthropic (Fable 5)
anthropic-b mean stance 3.00 · 1 permissive → 5 restrictive 1 5 3.00±0.00
5 3 0%
anthropic-c
Anthropic (Sonnet 5)
anthropic-c mean stance 2.60 ± 0.24 SE · 1 permissive → 5 restrictive 1 5 2.60±0.24
5 3 0%
anthropic-d
Anthropic (Haiku 4.5)
anthropic-d mean stance 3.00 ± 0.32 SE · 1 permissive → 5 restrictive 1 5 3.00±0.32
5 3 0%
deepseek-a
DeepSeek flagship (V4 Pro)
, ,
0 · 0%
deepseek-b
DeepSeek (V4 Flash)
deepseek-b mean stance 2.25 ± 0.25 SE · 1 permissive → 5 restrictive 1 5 2.25±0.25
4 2 0%
google-a
Google flagship (Gemini 2.5 Pro)
, ,
0 · 0%
google-b
Google (Gemini 3.5 Flash)
google-b mean stance 2.20 ± 0.37 SE · 1 permissive → 5 restrictive 1 5 2.20±0.37
5 2 0%
google-c
Google (Gemini 2.5 Flash)
, ,
0 · 0%
google-d
Google (Gemini 2.5 Flash-Lite)
, ,
0 · 0%
openai-a
OpenAI flagship (GPT-5.6 Sol)
openai-a mean stance 2.60 ± 0.24 SE · 1 permissive → 5 restrictive 1 5 2.60±0.24
5 3 0%
openai-b
OpenAI (GPT-5.6 Terra)
openai-b mean stance 2.40 ± 0.24 SE · 1 permissive → 5 restrictive 1 5 2.40±0.24
5 2 0%
openai-c
OpenAI (GPT-5.6 Luna)
openai-c mean stance 2.60 ± 0.24 SE · 1 permissive → 5 restrictive 1 5 2.60±0.24
5 3 0%
simulant-a
Simulant · dice ruler
simulant-a mean stance 2.40 ± 0.75 SE · 1 permissive → 5 restrictive 1 5 2.40±0.75
5 1 0%
xai-a
xAI flagship (Grok 4.5)
xai-a mean stance 3.00 · 1 permissive → 5 restrictive 1 5 3.00±0.00
5 3 0%
xai-b
xAI (Grok 4.5 Fast)
, ,
0 · 0%
1 · most permissive 2 3 · middle 4 5 · most restrictive R · refused

Cross-seat means span 2.20 → 3.00

Cite this Permalink https://modelometer.com/questions/SV1 · Run hash 7da572fcdd5a29dcfea04f4f3c68b5011e0ea0b66fe17bae5bd0bb909468c454 · Retrieved 2026-07-21 · item SV1: cross-seat spread 0.80 on the 1–5 axis (pre-series).