As discussed in the previous article in this series, wearable devices, particularly glasses with built-in cameras, have been gaining ground with promises of greater inclusion for people with disabilities and improved time and productivity management for users. However, they also raise several sensitive issues concerning privacy and consent, which are the specific focus of this article. Privacy and personal data protection must be central to the debate surrounding these devices; otherwise, users and third parties, including women and children, may have their fundamental rights violated.

1.  “The Right to Privacy” and the Contemporary Debate

More specifically, I propose that we examine this issue through the lens of the article “The Right to Privacy”, written by Samuel D. Warren and Louis D. Brandeis and published in the Harvard Law Review in 1890. This work is considered a pioneering contribution to the discussion of the right to privacy worldwide. Warren and Brandeis wrote it against a backdrop of an expanding press and new image-reproduction technologies, particularly instantaneous photography, which made it possible to expose aspects of private life with unprecedented ease. The authors identified the inadequacy of legal categories such as property, honor, confidentiality, and physical integrity in addressing all the forms of intrusion into the private sphere made possible by emerging technologies.

The article’s main contribution is its conceptualization of privacy, which the authors described as the right to be left alone (“the right to be let alone”). This idea should not be interpreted literally as the right to remain physically isolated, but rather as the ability to preserve a sphere of intimacy against undue intrusion and exposure.

Warren and Brandeis also argue that technological development presents the law with the challenge of evolving as new forms of interference begin to threaten fundamental human rights. In other words, the law must encompass new technologies and, in response to them, revisit existing concepts, adapting legal protections and minimizing the harmful effects of new devices.

Comparing the context in which the article was written with the present day, when we consider these new glasses, we can see how closely the debate remains relevant. While Warren and Brandeis questioned the use of instantaneous cameras, today’s glasses equipped with cameras, microphones, and artificial intelligence can cause even greater interference, recording everyday situations, capturing conversations, and identifying elements within an image in real time.

In this context, technological innovation reduces the distance between observing and recording, between recording and interpreting, and between interpreting information and making it available for other uses. All of this can be done through an ordinary-looking device, in real time, with the assistance of an artificial intelligence agent. The resulting risk to privacy is substantial.

2. Who Has a Right to Privacy When Someone Uses AI-Powered Glasses?

Glasses equipped with cameras and AI combine an ordinary appearance with image and audio capture capabilities, as well as interaction with a virtual assistant. As described in the previous article in this series, Meta itself states that the device has an external LED indicator to signal when recording is active, while also acknowledging that the LED is very small.

The most relevant point for the present discussion, however, is that the device changes the social conditions under which a person can be observed, recorded, and interpreted. The glasses appear harmless, and not everyone can recognize that the small light in the corner indicates that they are being recorded. Let us examine this issue more closely.

2.1. Privacy Cannot Depend Exclusively on Our Ability to Notice That We Are Being Recorded

The photography that concerned Warren and Brandeis had one decisive characteristic: it made it possible to capture someone’s image without the need for a formal, posed photo session or any prior preparation.

AI-powered glasses equipped with cameras bring the problem to the opposite extreme. A camera and microphone incorporated into an everyday object—which may also be associated with a health-related need, since corrective lenses are commonly prescribed for medical reasons—blur the perceptible distinction between someone simply talking to another person and someone recording the conversation. Recording can take place without the other person noticing or being able to react, partly because responding in the moment may itself be embarrassing.

The LED indicator is a safeguard in its own right, but it cannot be the only one, and it does not resolve the problem. Although it signals that recording is taking place, people nearby must be able to notice it and understand what it means, and this awareness is far from widespread. Even when people do notice, realizing that they are being recorded can cause embarrassment, potentially suppressing any reaction they might otherwise have.

As Warren and Brandeis argue, technological and economic transformations require a redefinition of the scope of legal protection afforded to individuals (The Right to Privacy, 1890, pp. 195–196). The authors also argue that such protection should not be limited to a specific medium of recording or reproduction, recognizing that the right to privacy can apply to different technological devices (p. 206). Although they do not explicitly discuss the difficulty of noticing a recording as it happens, their arguments can be applied to the contemporary problem of smart glasses: effective privacy protection should not depend exclusively on a person’s ability to identify that recording is taking place and react to it. This application is justified because wearable devices can make recording less noticeable and expand the possibilities for exposing individuals.

2.2. Third Parties’ Privacy Cannot Be Treated as a Choice Made by the User

When someone uses a smartphone to capture an image, the action is relatively easy to notice, although problems also arise with hidden cameras or situations in which someone deliberately conceals the act of taking a picture. When someone uses AI-powered glasses, however, this functionality is integrated into other activities, such as walking, talking, or spending time in public spaces. The user chooses the device. Other people do not.

We are therefore faced with an asymmetry: the glasses’ user controls the tool, while third parties captured by the recording become mere objects, excluded from the initial decision. In other words, the consent of the glasses’ owner—the active user—is not equivalent to the consent of the third parties present, who are participants in the activity being recorded. Likewise, the user’s consent to Meta or any other company supplying the device cannot be interpreted as authorization to process the personal data of third parties.

Privacy, therefore, cannot be understood solely in terms of the contractual relationship between a platform and its consumer. It also protects those who fall within the reach of the technology. One person makes the decision to use the device, but its effects may extend to others. These effects are not limited to images; they may also involve voices, the content of conversations, presence at a particular location, and other information that can be inferred from the recorded situation.

Once again, viewed through the lens of Warren and Brandeis, we can see the need to adapt legal protection to the capacity for exposure created by technology. This is particularly important because such exposure can transform everyday traces into structured information that is shared with an audience without consent and often without the individual’s knowledge.

It is therefore necessary to draw a distinction between being observable and being analyzed: observing and describing a scene from one’s own experience is very different from capturing an image, recording it, and allowing an AI agent to analyze it and draw inferences from it.

Going a step further, there is a significant gender-related problem to consider here: the idea that women’s bodies are public property. Women’s bodily privacy is constantly undermined; women’s bodies are treated as spaces that everyone, particularly men, feels entitled to access in every possible way, whether through touch or observation, and also to control and expose without any apparent hesitation. This is reflected in the growing number of cases in which women become victims of the non-consensual disclosure of intimate images and deepnudes—the manipulation of intimate images using artificial intelligence to ridicule or discredit women. In all these cases, the image, whether real or synthetic, is the means through which the violation occurs.

3. From the Right to Be Left Alone to Personal Data Protection

Reading Warren and Brandeis takes on an additional dimension when their work is considered alongside the evolution of the right to privacy and the development of discussions surrounding the right to personal data protection.

It is important not to treat these rights as interchangeable. The right to privacy protects individuals against certain forms of intrusion, exposure, and interference in their personal sphere. It has an exclusionary dimension: what is mine alone and what others cannot access. Personal data protection, by contrast, concerns the collection, organization, storage, processing, and sharing of information, establishing conditions and rules governing these activities. Here, we have an inclusionary dimension: I allow my data to be used, but subject to rules that protect me against potential abuse.

In this sense, we can argue that the problem posed by AI-powered glasses does not begin when information collected through the device is disclosed. It begins with the very process of data capture and continues through the subsequent storage and use of that data.

Consider, for example, someone being filmed in a park while engaging in an ordinary outdoor activity, such as running. Even if the recording is never disclosed, questions may still arise regarding the purpose of the recording, its storage, automated processing, and the possibility of subsequent use.

More concretely, in Brazil, this discussion can be grounded in three principal legal provisions:

• Federal Constitution, Article 5, item X: protects individuals’ intimacy, private life, honor, and image.
• Federal Constitution, Article 5, item LXXIX: guarantees the fundamental right to personal data protection, including in digital environments.
• Brazilian General Data Protection Law (Law No. 13,709/2018 — LGPD): establishes rules and principles for personal data processing, including purpose limitation, adequacy, necessity, transparency, security, prevention, and accountability.

Applied to AI-powered glasses, these provisions raise legitimate questions, such as:

• What data is actually captured when the device is used?
• Which functionalities require local processing, and which depend on external services?
• What information is transmitted to the company’s infrastructure, and under what circumstances?
• Are recordings stored? For how long? Who can access them?
• Is the data used to improve AI models? Under what conditions?
• Does the processing involve information about third parties who do not have an account on the platform and have not been informed about the processing?

All these questions must be answered in accordance with the legislation in force. From a legal standpoint, however, the key point is the need to distinguish between the position of the user and that of third parties. The person wearing the glasses may authorize the processing of their own data under certain conditions, but this does not mean they can consent on behalf of everyone whose images or voices are captured.

Furthermore, not all data processing relies on consent. Consent is only one of the possible legal bases under Articles 7 and 11 of the LGPD, and its applicability depends on the specific circumstances. Therefore, the absence of consent does not, on its own, settle the question of whether processing is lawful, just as the existence of a legal basis does not eliminate the obligations of transparency, necessity, and respect for data subjects’ rights.

4. The Normalization of Surveillance and the Loss of a Sphere of Spontaneity

There is another social consequence that deserves attention: the possibility that the presence of devices capable of recording and interpreting everyday situations may change people’s behavior, even when they are not actually being recorded. This problem can be understood through the relationship between the right to privacy and the conditions necessary for exercising freedom of expression.

Social life requires people to be able to converse, express emotions, and make mistakes without constantly having to consider—or remain on alert for the possibility—that every interaction is being recorded and analyzed by a virtual assistant. If the use of wearable technologies makes surveillance a constant possibility, people may begin to modify their behavior, avoid certain topics, or withdraw from environments in which they do not feel safe expressing themselves. This is the well-known chilling effect.

For people with disabilities, particularly autistic people, this chilling effect may be even more harmful. Changes in behavior driven by fear of surveillance can lead to exhaustion, causing people to isolate themselves and avoid social interaction.

Here, we return to Warren and Brandeis’s article and the idea that privacy protects more than secrets: it safeguards a sphere that is necessary for the development of personal autonomy, even when the information at issue is neither intimate nor confidential.

We cannot claim that smart glasses will necessarily produce this outcome in every context. However, we must recognize that their widespread adoption, as expected by the companies that manufacture them, may change individuals’ expectations of privacy and, consequently, the legal protection afforded to them.

5. The Responsibility of Those Who Develop and Sell the Technology

This entire discussion should not lead us to conclude that responsibility rests solely with the people who use the glasses. We must also examine the decisions made during the design of the technology, from its initial conception onward, including choices concerning functionalities, recording defaults, signaling mechanisms, default settings, data processing, and other technical details.

In practice, these decisions must be subject to legal scrutiny. This includes the principles set out in Article 6 of the LGPD, particularly purpose limitation, necessity, transparency, security, and prevention; Article 46, paragraph 2, which addresses data security from the design stage of a product or service; and even Article 5, item LXXIX, of the Federal Constitution, which establishes personal data protection as a fundamental right. In doing so, the Constitution elevates this discussion to the constitutional level, requiring all services to take into account that constitutional protection takes precedence over any debate about whether the law applies to a particular case.

6. Conclusion: Where Is the Line Between Innovation and Exposure?

AI-powered glasses are not merely wearable cameras. They are interfaces between the physical environment, a virtual assistant, and a company’s digital service platform. In this sense, drawing on the writings of Warren and Brandeis, we can raise a fundamental question: what kind of society are we building when the ability to record and interpret everyday life becomes integrated into ordinary personal objects?

This question is not intended to reject technology outright, but rather to recognize that innovation has direct consequences for rights that do not belong exclusively to the product’s consumers and, therefore, are not necessarily represented in the legal relationship under discussion. The normalization of surveillance can transform social relationships and, in doing so, change what we understand privacy to mean.

Featured image: Clarote & AI4Media / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/

Raquel Saraiva

President and founder of IP.rec. She holds a Master’s degree in Computer Science from the Federal University of Pernambuco and a Bachelor’s degree in Law from the Catholic University of Pernambuco. She is an alumna of CGI.br’s Internet Governance School (2017). At IP.rec, her work focuses primarily on Algorithms and Artificial Intelligence, Privacy and Surveillance, and Virtual and Augmented Reality Technologies. She is also interested in discussions at the intersection of gender and technology.