“A great step forward for respect for privacy in AI.” According to this prayer, Apple promised a rarity: respecting the personal data of the users, in this case, those who wish to take advantage of generative artificial intelligence in iPhone, iPad and Mac.
A little less than a year later, a researcher shows that Apple’s promises are far away, far from being held.
Confidential data sent to Apple
As reported by the Dark Reading Specialized Site, the results of Yoav Magid’s research, a Lumia Security security researcher, shows that Apple recovers more or less personal and contextual information to operate Apple’s intelligence.
If you ask Siri while Siri can recover and send the company’s servers to the music that I hear at the same time. More worrying, if you use Siri to send a message through an encrypted application, such as WhatsApp or Imessage. The content of the message and contact information are sent to Apple servers.
To understand, you must first know that Apple Intelligence works in two ways. Certain tasks can be performed locally, directly on the user’s device. While others, which require more computer power, are administered by remote servers of what Apple describes as ultra insurance and calls the private cloud.
Ai and the myth of privacy?
These data transmissions are worrisome about the privacy of Apple’s intelligence users, and obviously require a clear and precise response from the engineers of the US giant, in particular to know what these additional data can serve. We have also contacted Apple to return to these questions, obviously we will update this article as soon as the case.
Be that as it may, according to the cybersecurity expert, the way Apple has managed this revelation is worrying, it is not surprising, in view of the relationships sometimes a bit tumultuous that the giant has with security researchers.
Contacted by Yoav Magid, Apple first collaborated with the researcher and showed interest in his discoveries. Then, the representatives of the Cupertino firm minimized their findings and responded that they were “expected behaviors” and that this way of doing fulfills their confidentiality letter. Rarely consulted in users. “It’s surprising because Apple emphasizes privacy, but collects so many data,” he explains.
This public made report, above all, makes it possible to take into account that with AI, and even more with technological giants, does it not save privacy, or perhaps it is not a possibility? To do? Yoav Magid explains how it proceeds when using “intelligent” services. Without reading the conditions of use and respect for privacy, systematically deactivates the characteristics that allow the learning of uses. It is not a miraculous remedy, but it is already a good start. The myth of information collection to “improve services and products” is only there to recover better personal information and use statistics.
Source: BFM TV
