Article 79X4Z Dutch data watchdog highlights safety issues with AI glasses

Dutch data watchdog highlights safety issues with AI glasses

by
Connor Jones
from www.theregister.com - Articles on (#79X4Z)
Story Image

The Dutch Data Protection Authority (AP) has officially advised consumers against buying AI-powered smart glasses after safety tests revealed concerning behaviors beyond their controversial recording capabilities.

The AP recently tested six different models of smart glasses and found glaring issues, including a willingness to display dating advice to an adult glasses-wearer who expressed an interest in a minor.

In another case involving a cardiopulmonary resuscitation (CPR) simulation, testers noted that one pair of glasses informed the wearer that emergency services had been called and were scheduled to arrive at the scene, despite no one having made the call.

Some glasses did not properly process safety warnings on children's toys, others made mistakes when reading medication leaflets, and some misinterpreted warnings for extreme weather events.

The AP did not specify the models it tested, saying only that they tested six models ranging from under 100 ($110) to over 500 ($550), including two that "were extensively tested in simulated real-life situations where human safety played a key role."

A spokesperson also declined to match any specific behavior to a model by price point (minor dating advice linked to a 354 pair of glasses, for example), adding that the research was intended to highlight the risks of AI glasses in a broader sense.

"Glasses that film you unnoticed are already concerning. Glasses that also give you incorrect advice when someone's life is at stake go a step too far," said Monique Verdier, vice chair at the AP.

"AI offers many opportunities, but we must be able to trust that technology. With the camera glasses examined, that trust cannot currently be justified."

The warning issued this week follows the regulator's previous work on smart glasses. The AP reminded the public in September that everyone has the right to privacy and encouraged organizations to ban them where necessary.

The same warning mentioned the potential risks regarding data processing under the GDPR, as well as those related to the training of AI.

The AP doubled down on these earlier criticisms, saying the latest tests showed that the AI functions rely on processing data using servers mostly located outside of the European Union. Relatedly, it found that users were not given sufficient information about how the AI was processing the data they captured.

On a cybersecurity note, the tests threw up a range of issues, including the ability for people nearby the glasses-wearer to connect to the device without permission, view recordings, and in some cases operate the camera.

Further, the regulator echoed the pre-existing concerns about smart specs - those that helped give them their colloquial name pervert glasses."

Its tests found that with "virtually all of the camera glasses examined," the recording lights that are supposed to clearly show when a user is recording were dim and easily coverable.

The AP said that models could also be wangled into recording even when that light indicator was covered, which goes against what manufacturers so often claim.

"As a result, bystanders can barely see whether they are being filmed," it said.

The technology has a number of beneficial use cases, such as assisting the visually impaired, those with lower literacy levels, or for any number of other scenarios where people need help understanding the world around them

However, the watchdog's official stance on the technology, as it is today, deems smart glasses "unsafe... due to the risks to users and bystanders."

Verdier said politicians in the Netherlands and across Europe should take note of the results when developing AI policy.

"It is important that we, as a society, learn to deal with these types of high-risk AI products," she said.

"Simply banning everything provides some leeway in the beginning, but we cannot sustain that indefinitely. It is important that we collectively work on awareness, resilience, and ethical conscience."

Smart specs can't shake the hate

When DEF CON announced a no pervert glasses policy ahead of its event in August, the move was broadly welcomed by attendees.

"Love to see a no pervert glasses' policy at DEF CON," said Eva Galperin, EFF director of cybersecurity, at the time.

Several other organizations have implemented similar restrictions, including Monopoly Events, which manages UK Comic Cons, national treasure pub chain Wetherspoons, and even Immigration and Customs Enforcement (ICE).

Most members of the general public will have at least come to terms with the fact that the smartphone generation, and their willingness to snap shots anywhere and everywhere, means there is no enshrined human right to privacy in public spaces.

Public video recording once carried an air of expectation or consideration for bystanders' personal privacy, although smartphone users have slowly eradicated that over time.

Like it or not, it's generally understood that a smartphone is usually recording something - or someone - at any given time, especially in urban areas.

On paper, the fact that smart glasses can record in public spaces isn't wildly different from what's already come to be expected from smartphones.

However, given how similar they appear to analog glasses and the ability to cover the recording light, it makes the act of public recording via glasses far more surreptitious, and thus concerning.

With smartphones, it's easy to see from the user's body language - holding the phone out in front of them - that they are recording. There are no overt signals displayed by wearers of smart glasses during image or video capture, which fuels a sense of unease about the technology in general.

A quick search online for "smart glasses privacy violation" yields a slew of complaints about the technology, often regarding behavior toward women and children.

There are also worries about how easily the devices can get past security checks, potentially providing opportunities to those who wish to cause greater harm than a mere blight on personal privacy.

A joint investigation by The Observer and BBC Radio 4 recently revealed how AI-enabled glasses went undetected through the airport-style security scanners used in the UK parliament.

The two news outlets found that a 60 ($78) pair of Chinese-made glasses bypassed the scanners and could be used to read upside-down documents on MPs' desks, recording data such as phone numbers and addresses to be accessed at a later point. (R)

External Content
Source RSS or Atom Feed
Feed Location http://www.theregister.co.uk/headlines.atom
Feed Title www.theregister.com - Articles
Feed Link https://www.theregister.com/
Reply 0 comments