As AI-equipped electronic devices rapidly enter everyday life, backlash over privacy infringements is growing, IT outlet TechRadar pointed out on Aug. 8 (local time). TVs with background recording, earbuds that offer call recording and smart glasses that can film what wearers see are now easy to find. In an era when glasses watch, earbuds listen and TVs get too deeply involved in our lives, how far should consumers allow technology to intrude?

Lauren Hendry Parsons (로런 헨드리 파슨스), communications director at the Mozilla Foundation, saw this as not a simple new-product controversy but a bigger trend toward normalising consumer surveillance. In the past, surveillance devices stayed in the form of smart speakers in the home or fixed IoT devices. Now they have shifted to mobile forms attached to the body, such as earbuds and glasses. As in cases where fashion influencers use Ring doorbells like tools to film an outfit of the day, surveillance is increasingly being packaged as a stylish lifestyle. Parsons voiced concern, calling it the “chicification of surveillance.”

The core issue is who is considered a user. Buyers of smart glasses or earbuds can control the functions. But people sitting at the same table, call counterparts, children and passengers are also subject to data collection while lacking control. Parsons stressed: “True ‘Privacy By Design’ is protecting not only device purchasers but also people around them who unintentionally enter a collection environment.”

Parsons proposed surprise-free design, meaningful user control, limited data collection, reasonable default settings and layered protection as solutions. “A notification is not consent,” she said, stressing that a short beep indicating recording has started or a legal notice alone is not enough. The more sensitive a function is, the more proportionate friction there should be in the consent process, she explained.

The burden of the problem does not fall equally on everyone. “The risks of ambient recording and filming weigh more heavily on people who already have less control over their situations, such as workers, children, parents, patients, tenants, domestic violence victims and protesters,” Parsons said. Surveillance may look like a convenience feature, but it can in practice make vulnerable groups more vulnerable, she added.

A different direction for AI was also suggested. Div Garg (디브 가르그), co-founder and chief executive officer of AGI, said: “AI wearables do not necessarily have to send data to the cloud.” If models run on the device, there is no need for data to leave it, and users can own the model and the data, he argued. Garg stressed that AI should be designed not as a data collection tool for advertising revenue but in a neutral way that truly helps users.

Garg described this as a “Zero-Trust” approach. Without needing to trust the provider, if it is ensured that the AI model works only on the device, the very structure in which data does not leave the device becomes the basis for trust. For that, he said, technological innovation is needed to make models smaller and optimise them to run on smartphone GPUs and NPUs without battery drain or overheating.

Calls are growing to make consent and privacy core requirements in product design rather than simply arguing to ban surveillance-style AI devices. “The future of technology is not inevitable,” Parsons said. “Better choices by companies, strong regulatory standards and appropriate friction can protect people.” Garg also said he expects the future will require not the same chatbot for everyone, but AI assistants that work inside each person’s device and are tailored to the individual.