Facebook has admitted to hiring human contractors to listen to Messenger conversations just weeks after Google was discovered to potentially be breaching the EU’s General Data Privacy Regulation (GDPR).
The Menlo Park-headquartered social media giant admitted that contractors have been able to listen in on and transcribe conversations held on the Messenger app. The news was first reported by Bloomberg. The contractors were hired to make sure that the transcription function was working properly. Facebook claimed that it had stopped workers from reviewing audio clips “more than a week ago”.
While the transcription service is switched off by default, if anyone else taking part of the conversation has switched on the service then the chat may be listened to and transcribed, according to The Verge.
The report comes after Google, Apple, Microsoft and Amazon have all been reported to using human contractors to listen to audio obtained through their voice assistant products without being explicitly transparent with their users.
RegTech Analyst recently reported on how the Hamburg Commissioner for Data Protection and Freedom of Information (HmbBfDI) has launched an investigation against the tech titan for potentially failing to live up to the standards set out by GDPR.
Just like in Facebook’s case, not everyone who were recorded were aware that they were. “Privacy professionals often caution against the risk of opaqueness in using AI systems to process personal data,” Emma Erskine-Fox, associate at TLT, the UK law firm, tells RegTech Analyst. “There’s no doubt that it’s crucial that businesses must be transparent about their use of AI where that use affects individuals. However, the recent admission by several big tech companies that humans have been listening to users’ voice recordings without those users’ knowledge shows that it’s just as important to be transparent about where and how humans are involved in ostensibly automated processes.
“Individuals may reveal more about themselves if they think they are only talking to a robot, leading to potentially significant privacy ramifications if other humans have access to sensitive data without users’ knowledge. Therefore, there are some important questions that must be answered: is there a trade-off being made between transparency and accuracy of AI output? What are the ethical implications of involving human intervention without users’ knowledge or consent? How can outputs be effectively audited whilst complying with fair processing obligations under data protection laws?”
The news of contractors listening in on people’s Messenger conversation also comes as Facebook is facing increased scrutiny over its new cryptocurrency Libra. In early August, several international authorities asked the social media company to be more open about the initiative. They were particularly concerned about potential privacy risks associated with Libra. The institutions included organizations from the UK, the US, Canada, Albania and Australia.
Copyright © 2018 RegTech Analyst
