Robot Surveillance Contracts

We are emerging into a world where all manner of automated systems engage in surveillance. Internet spiders (bots) collect intelligence from our web sites and servers. Search engines record our Internet queries. Security cameras surveil our bodies and our vehicles as they move about in physical space. Smart phones keep detailed records about incoming transmissions from third parties. Mobile apps track our GPS location.  Motorists mount cameras on their dashboards to memorialize what happens in the event of a traffic collision. The day is dawning when robots record their encounters with people, businesses and property.

The trend toward increasing robotic reconnaissance is unmistakable. The trend will continue to grow on account of digital technology and our legal system's veneration of records. Technology makes the collection and analysis of text, image, audio, video and other forensic records ever more easy. At the same time, people want to possess those records for legal purposes. Our legal traditions hold records in high esteem, as they are powerful agents in the resolution of disputes and the assignment of accountability.

But just as our legal system honors records, it also has a tradition for honoring agreements – that is, contracts. When two parties communicate, the legal system respects the agreements they make between themselves. Contracts can even be formed through the communication of machines, such as computers. And, contracts can cover topics like the creation, use, protection and disposal of records.

Implication: contracts can establish terms to limit and regulate automated surveillance.

Thus, a web administrator can post "terms of service" on its site, and the legal system will (generally) uphold those terms as binding on automated spiders that probe the administrator's site. (For example, Internet Archive and Suzanne Shell settled a lawsuit where Shell claimed that IA copied material from her web site in violation of her published terms of service. Instead of pressing forward for victory in the lawsuit, the Internet Archive settled. It publicly expressed regret for encroaching on Shell’s rights. See also Mark Rasch's discussion of web terms-of-services cases.)

The general principle is this: People (and businesses) can post or broadcast legal terms that impact automated snoops. The terms are like an end user license agreement (EULA) associated with software.

For this general principle, I foresee a large application. People can communicate all kinds of terms to and form contracts with robots.

Here is an example of legal terms a person could publish: "Notice to any system that may be surveilling my activities or entering the domain of my systems or probing or querying my systems: You do not have permission to make or store records about me, my activities, my systems, the contents of my systems or communications from my systems. By making and storing such records, you agree as follows: You will treat me with dignity. You will destroy the records immediately. If you violate your agreement and do not destroy the records, then you agree to keep the records confidential, you agree not to use them in a way that is contrary to my security or my interests and you agree not use them without my written permission. If you do not destroy the records, you also agree to e-mail within one day of creation a copy of all such records to me at You agree that these terms may be enforced by judicial injunction, together with any other appropriate legal remedy, including monetary damages."

Terms like this could be published on a web page. Or they might be transmitted to a spy's e-mail address. Or they might be posted for visual reading on a vehicle or a building. Or they might be transmitted via an open Bluetooth channel (or multiple radio frequencies) so that a local robot, surveillance camera or store owner could detect them.  Or they might be transmitted by way of an augmented reality channel detectable by a local robot or camera.

The publication of such terms is like posting an end user license agreement around yourself or your property.

Is there any reason why terms like this should not be legally binding on nosey systems or robots that can access the terms?

[Nothing I publish publicly is legal advice, but the foregoing is something to think about and debate in public.]

1 comment:

  1. Imagine a homeowner who did not want Google StreetView photographing her house, her dog and her pansies planted in the front yard. She could post a sign in her yard, facing the street -- a form of end user license agreement -- stating that she forbade search engines from photographing her property and placing the results in an online system. Her sign could say that by photographing the property a search engine showed its agreement to the terms posted on the sign. Whether the sign would hold up in any particular court is an interesting question. But as a practical matter, I'll bet that if this sign were particularly brought to Google's attention, it would just wipe her property off of its records. [No one should take my ideas as legal advice, just something to think about.]