HOW TECH COMPANIES ARE NOW
ADDRESSING THESE PRIVACY HOLES
Apple initially responded to the Guardian’s
story with: “A small portion of Siri requests are
analyzed to improve Siri and dictation. User
requests are not associated with the user’s
Apple ID.” However, days later, Apple opted to
globally suspend using human contractors for
the analysis of Siri commands. In a statement
to The Verge, the Cupertino company
explained: “Additionally, as part of a future
software update, users will have the ability to
choose to participate in grading.”
Similarly, Google has temporarily halted
its practice of transcribing speech data from
Google Assistant in the European Union,
where it is suspected that the assistant could
have been violating the EU’s GDPR laws
pertaining to data protection. Microsoft’s
reaction, meanwhile, has been more subtle;
it has simply changed its privacy policy to
openly acknowledge that, yes, humans could
end up processing recordings of the utterances
you make to Cortana.
Will all of these scandals – or mini-scandals,
you could call them – prompt tech companies
to significantly lessen their reliance on
human analysis of users’ voice recordings?
The short answer is that it looks unlikely. As
Macworld staff writer Jason Cross explains,
it’s impossible to train machine-learning
algorithms without a human element. “If there
was a computer algorithm that could always
accurately determine whether the AI was right
or wrong, it would be the AI algorithm!”