Meta in row after workers who saw smart glasses users having sex lose jobs
Meta’s Ray‑Ban smart glasses are under fire after outsourced workers said they were routinely exposed to intimate footage, including people having sex, while labeling videos sent to Meta’s servers. Commenters question why any private recordings from wearable cameras are reviewed by humans at all, highlight the lack of consent for bystanders and partners who are filmed, and link this to broader concerns about non‑consensual imagery, CSAM moderation, and the offloading of traumatic AI‑training work to low‑paid contractors. Many see the episode as another example of Meta’s privacy‑hostile incentives and argue for stricter norms or regulation around always‑on cameras in consumer devices.
Scope of the discussion
- Focus on Meta’s smart glasses uploading user video (including nudity/sex) to Meta, where low‑paid contractors review/label it, and Meta then ending its contract with the Kenyan outsourcer after workers spoke to the press.
- Many comments treat the news as confirmation of long‑standing worries about Meta’s attitude to privacy and data exploitation.
Smart glasses vs. other cameras
- Some argue smart glasses are no worse than smartphones or CCTV: you’re already recorded constantly in public.
- Others see glasses as categorically different:
- Always pointed where you look, easy to mistake for normal eyewear.
- Recording indicators can be hidden; much better for covert creepshots/NCII than a conspicuous phone or GoPro.
- Harder to socially or institutionally enforce “no recording devices” when they look like prescription glasses.
- A minority defend benign uses (cycling, POV sports, accessibility for blind users) but even they often say they wouldn’t trust Meta to run them.
Meta’s data practices and consent
- Many take it as self‑evident that Meta designed the system to harvest and label intimate footage and that this is unacceptable, regardless of TOS.
- Others note users technically “consented” in privacy policies allowing sharing with service providers for analysis and product improvement.
- Strong pushback: legal click‑through consent is framed as manipulative and not meaningful moral consent, especially for bystanders who never agreed to be filmed.
- Former Meta employees describe strict internal controls on employee access to user data; others counter that contracted moderators are outside those safeguards and that leadership will override rules when it suits monetization.
Whistleblowing, contractors, and trauma
- Some say Meta dropping Sama is predictable: big companies won’t tolerate vendors generating bad press.
- Others emphasize the ethical duty to expose serious privacy abuses, regardless of contractual fallout.
- Broader criticism of “trust & safety” outsourcing:
- Moderators in poorer countries are exposed to disturbing content (including CSAM) for very low pay and minimal psychological support.
- Debate over whether such work can ever be ethically acceptable and what pay/support would be required.
CSAM, scale, and platform design
- Large subthread on whether platforms at Meta’s scale can ever moderate CSAM adequately.
- Proposals:
- Federated, small servers with legally responsible moderators to shrink reach and improve responsiveness.
- Counterarguments:
- Decentralization could simply push CSAM into poorly moderated or complicit servers, increasing total volume and making enforcement harder.
- General agreement that current approaches by big platforms and governments are inadequate.
Regulation, norms, and Meta’s reputation
- Mixed views on bans:
- Some want outright bans on camera‑glasses (at least in certain spaces); others warn this would hurt accessibility and legitimate uses.
- Strong distrust of Meta:
- Seen as structurally committed to maximizing surveillance and engagement, not user welfare.
- Many say they would never buy Meta hardware with a camera or already avoid all Meta products.
- Pessimism that this incident will materially change user behavior, regulation, or Meta’s conduct.