Build1 publisher3 min readPublished
Companies are asking France's CNIL whether they can bar smart glasses from their premises
Paris prosecutors have opened at least one criminal case over street videos of women filmed without consent. The data regulator has fewer than ten workplace complaints on file and a growing queue of employers asking about bans.
The Engineer · Build desk
What happened
- The Paris prosecutor's office opened at least one criminal proceeding after sexual harassment complaints tied to videos of women filmed in the street without consent and then posted online.
- The prosecutor's cybercrime unit tested smart glasses two weeks before the report, to work out which offences the devices make possible.
- France's data protection authority CNIL has received fewer than ten complaints about smart glasses being used in workplaces.
- A growing number of companies are asking CNIL whether they are allowed to forbid employees from using smart glasses on company premises.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- constraint An employer permitting the glasses cannot later produce evidence that anyone in the room was informed, because the device's only notice to bystanders is a light on the frame.
- decision Neither the prosecutors nor the regulator has named a brand, so a workable site policy has to be written around any eyewear containing a camera, and whoever staffs the entrance has to enforce it.
- exposure The criminal exposure described in the record sits on the person doing the filming, at up to a year in prison or 45,000 euros, and the report describes no action against an employer.
- precedent If Australia bans the glasses on its own government sites, private employers get a published rule to copy while CNIL's answer is still pending.
The camera is built into the frame, and the report describes it as able to record without drawing attention [9]. Notice to everyone else is a small indicator on the frame that lights when recording begins [18]. The person being filmed acknowledges nothing, and the light goes unrecorded.
CNIL frames the problem the same way. Nassera Bekhat, head of economic affairs at CNIL, said smart glasses make it genuinely difficult to inform people that they are being filmed and to obtain consent where consent is required, quoted in translation in mezha.net's account of Reuters reporting [14][26]. Devices with AI features fall under the EU AI Act and GDPR, which set the rules for handling personal data [15].
Rights groups say the indicator light moves the work to the person being filmed, who has to notice it and get out of the way [19]. Ines Legendre, a lawyer at the online safety organisation e-Enfance, said the light coming on does not mean the person consented to being filmed or to their image being distributed to thousands of people [20]. The same groups say the abuse mostly targets women [22].
Anyone drafting a site rule hits a scoping problem first. Paris prosecutors have not disclosed which glasses were used in the case they opened [3], and CNIL has not named the brands behind its complaints either [6]. So the rule has to cover eyewear with a camera in it, and enforcing that at an entrance means asking security staff to identify, at a glance, the one pair of spectacles in the lobby that is recording [9].
Meta sold 7 million AI glasses during 2025 [10], and its devices with EssilorLuxottica hold roughly 76% of the global market [8]. Dividing one by the other puts the whole market near 9.2 million units [25]. For that to be a unit count, the 76% would have to be a share of units over the same year. The report does not state the basis. EssilorLuxottica's most recent figures point to exponential growth in 2026 [11].
The criminal route is narrower than the workplace one. French law covers recording a person's image in a private place without consent, with a maximum of one year in prison or a fine of 45,000 euros [16], and enforcement gets harder once the video is circulating online [17]. Whether an office or a shop floor is a private place, the report does not say.
Meta answered on the product. A Meta representative said the company built privacy into its AI glasses from the start and will keep strengthening protections as the capabilities of the glasses expand [12]. EssilorLuxottica declined to comment [13].
Australia said on 17 September that it is considering barring camera smart glasses from government institutions over privacy and security risks, and suggested it would be the first ban of its kind [23][24]. A French employer writing the same rule now does it before CNIL has answered the questions already on its desk [7].
What to watch
- Whether CNIL publishes a position on employer bans, and whether it addresses premises-wide prohibitions rather than case-by-case complaints.
- Whether the Paris case identifies a device model or leads to charges. Either would show employers what evidence a prosecution actually needs.
- Whether Australia adopts its government-premises ban and how it defines a camera-equipped pair of glasses.