United StatesTechnology and telecom2016 – ongoingRing
Customer-video labeling for computer-vision training
The FTC proposed a $5.8 million Ring settlement covering unconsented video use and security failures
Problem
Ring used camera recordings to improve its products. Training algorithms involved people reviewing footage and labeling people or objects in customers' private spaces.
Approach
The FTC alleged that Ring gave employees and contractors excessive access to recordings without adequately informing customers or obtaining consent. Its complaint said Ring began restricting research footage to publicly posted or expressly consented recordings in January 2018. A proposed 2023 order required deletion of affected recordings and derived models or algorithms.
Results
Takeaway
The proposed remedy reached models derived from improperly used recordings. Restricting future access alone did not address the earlier training data.
Categories
- Size
- Large enterprise
- Industry
- Technology and telecom
- Business function
- IT and R&DFinance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Built in-house
- Goal
- Quality and risk
- Evidence
- Regulator, audit or court
Sources
Source summaries
The FTC alleged unauthorized use of customer videos for algorithm training alongside broader security failures. Its proposed order included deletion of derived models and a $5.8 million payment for refunds, subject to court approval.
The complaint described human labeling of people and objects for algorithm training. It said Ring began limiting research and development footage to public recordings or recordings with written consent in January 2018.
Full details are in the original sources.