⚠️ Ethical & Safety Concern: Human Gesture Detection via Signal Processing May Pose Unintended Risks

[unrealandychan]

⚠️ Ethical & Safety Concern: Human Gesture Detection via Signal Processing May Pose Unintended Risks

Summary

I am raising this issue to highlight a serious ethical and safety concern regarding the use of signal-based human gesture detection in this project. While the technology is impressive and the engineering is commendable, I believe the community should openly discuss the potential implications of using signal processing (e.g., RF signals, radar, Wi-Fi CSI, or similar sensor modalities) to detect and interpret human body gestures — particularly around privacy, consent, misuse, and long-term societal impact.

---

Background

RuView appears to leverage signal-based sensing to detect human gestures without requiring cameras or direct physical contact. This is a fascinating area of research, but it sits at a crossroads of several sensitive domains:

• Passive surveillance — signals can penetrate walls and detect people without their knowledge
• Biometric profiling — gesture patterns can be unique to individuals, functioning as a biometric identifier
• Behavioral inference — extended gesture monitoring can reveal health conditions, emotional states, or behavioral patterns

---

Concerns

1. 🕵️ Non-Consensual Surveillance

Unlike cameras, signal-based gesture detection can operate through walls and without any visible hardware. A person in an adjacent room, or even a neighboring apartment, could be monitored without ever knowing. This raises profound questions:

• Is there a mechanism in place to ensure informed consent from all individuals in the sensing range?
• How does the system define and enforce a sensing boundary?
• Could this be deployed covertly by a malicious actor?

2. 🧬 Biometric Data & Privacy Regulations

Gesture patterns, gait, and movement signatures can be biometrically identifying. Depending on jurisdiction, this may fall under:

• GDPR (EU) — biometric data is classified as sensitive personal data under Article 9
• CCPA (California) — biometric identifiers are explicitly protected
• BIPA (Illinois) — strict regulations on collection and storage of biometric data

Does RuView currently include any guidance on data retention, anonymization, or compliance with these regulations?

3. 🧠 Health & Psychological Inferences

With sufficient data, gesture-based signals can potentially infer:

• Neurological conditions (e.g., tremors indicative of Parkinson's disease)
• Mental health states (e.g., agitation, anxiety patterns)
• Sleep behavior and disturbances

This level of inference goes far beyond simple gesture recognition and enters deeply sensitive medical territory — without any of the safeguards that medical devices are required to uphold.

4. 🎯 Potential for Misuse & Weaponization

Signal-based gesture detection at scale could be misused by:

• Governments or authoritarian regimes for covert population monitoring
• Employers to surveil remote workers without their awareness
• Stalkers or abusers to monitor individuals in their homes
• Commercial actors to profile consumers in retail or public spaces without consent

The open-source nature of this project, while valuable for research, lowers the barrier for such misuse significantly.

5. ⚖️ Lack of Ethical Framework

There does not appear to be any documented ethical use policy, responsible disclosure framework, or contributor code of conduct specifically addressing the dual-use nature of this technology. Most cutting-edge sensing technologies today are expected to ship with:

• An ethics statement
• Use-case restrictions
• Guidance for researchers and developers on responsible deployment

---

Recommendations

I would like to respectfully propose the following steps to address these concerns:

1. Add an Ethics & Responsible Use document (ETHICS.md) to the repository that clearly outlines acceptable and unacceptable use cases.

2. Implement consent-aware design principles — document how developers using this library should obtain and manage user consent.

3. Add a disclaimer in the README.md clearly stating the legal and ethical obligations of anyone deploying this technology.

4. Scope-limiting guidance — provide technical recommendations to limit sensing range so that it does not extend beyond intended boundaries (e.g., into neighboring private spaces).

5. Data minimization by default — ensure raw signal data is not stored beyond immediate inference needs, and document this in the architecture.

6. Engage with an ethics review board or advisory panel — consider partnering with privacy advocacy organizations or academic ethicists to audit the project.

7. License restrictions — consider adopting a license such as the Responsible AI License (RAIL) that prohibits harmful or non-consensual use cases.

---

Conclusion

Signal-based human gesture detection is a powerful and genuinely exciting technology. But with great power comes great responsibility. The same capability that enables a smart home to respond to a wave of your hand could — in the wrong hands — enable pervasive, invisible, and non-consensual surveillance of human beings in their most private spaces.

I raise this not to discourage the work, but to encourage the maintainers and contributors to take proactive, concrete steps toward responsible development. The open-source community has a unique opportunity — and obligation — to set a high ethical bar before this technology is widely deployed.

I am happy to contribute to drafting any of the above documents if the maintainers are open to collaboration.

Thank you for your time and consideration.

---

Labels suggested: ethics, privacy, security, discussion, documentation
Type: Concern / RFC (Request for Comment)

[mfl22]

This was closed by authors without response? Ridiculous, as are their "comments" about "privacy-preserving" in their paper.

[ruvnet]

Thank you for raising this. I appreciate the thoughtful discussion and I agree that technologies like this deserve careful consideration.

I built RuView to explore a different approach to sensing people in environments where cameras are either intrusive or inappropriate. Many of the use cases I care about are things like fall detection, overdose monitoring in bathrooms, sleep monitoring, and general presence awareness in spaces where video would feel invasive.

The system works by analyzing disturbances in radio signals such as WiFi CSI or radar reflections. It is important to understand that this is not high resolution imaging. It is noisy signal data that requires calibration, proximity, and cooperative hardware placement. In practice the sensing range is limited and the fidelity drops off quickly outside the intended environment.

That said, the broader point about responsible use is valid. Any sensing technology can be misused if deployed irresponsibly. Cameras, microphones, and even basic WiFi infrastructure carry similar risks. I believe the right response is transparency and guidance rather than pretending those risks do not exist.

I am open to adding an ETHICS.md and responsible use guidance to the repository. My intention with this project is to explore privacy preserving sensing that can replace cameras in many situations, not create tools for covert monitoring.

If you are interested, I would welcome help drafting a responsible use section that outlines consent, transparency in deployment, and data minimization principles.

Appreciate you taking the time to raise the discussion.

[krzysztofantczak]

@unrealandychan what is your real end goal here? Are you assuming that some file will prevent bad actor from misusing the tech? How did that helped AI in general? How does it help progress? How does it help anyone - besides people that think their conscience is clean because they drafted some text file? This is not rage bait, I'm actually curious on what people think will happen.

@ruvnet I'm in the process of testing this thing on my hardware. Is it even possible to utilize it from outside of your house without being connected to your wifi? I might be wrong, but I don't think so? And if that's the case, why is it any different from your in-house security cameras (in terms of safety concerns).

[ruvnet]

As for my goal, I don't have any goal, I built it on whim on my Live Cast months ago, forgot about it, it took off. Now that's it out there, I suppose it's my responsibility to provide some guide rails. I didn't include the firmware initially and everyone said it was fake, which literally involved asking Claude to create some firmware. but i suppose most people didn't figure this out and assumed it was fake etc. So i created v2, mostly to prove the doubters wrong..

A bit more detail: https://x.com/ruv/status/2029187274081026280