on the Use of RFID on Consumer Products
Consumers Against Supermarket Privacy Invasion and Numbering (CASPIAN)
Privacy Rights Clearinghouse
American Civil Liberties Union (ACLU)
|American Council on Consumer Awareness, Inc.
Center for Democracy and Technology (CDT)
Consumer Project on Technology
Simson Garfinkel, Author, Database Nation
Edward Hasbrouck, Author, The Practical Nomad
Massachusetts Consumers' Coalition
|National Association of Consumer Agency Associates
Private Citizen, Inc.
Virginia Rezmierski, Ph.D.
World Privacy Forum
Radio Frequency Identification (RFID) is an item-tagging technology with profound societal implications. Used improperly, RFID has the potential to jeopardize consumer privacy, reduce or eliminate purchasing anonymity, and threaten civil liberties.
As organizations and individuals committed to the protection of privacy and civil liberties, we have come together to issue this statement on the deployment of RFID in the consumer environment. In the following pages, we describe the technology and its uses, define the risks, and discuss potential public policy approaches to mitigate the problems we raise.
RFID tags are tiny computer chips connected to miniature antennae that can be affixed to physical objects. In the most commonly touted applications of RFID, the microchip contains an Electronic Product Code (EPC) with sufficient capacity to provide unique identifiers for all items produced worldwide. When an RFID reader emits a radio signal, tags in the vicinity respond by transmitting their stored data to the reader. With passive (battery-less) RFID tags, read-range can vary from less than an inch to 20-30 feet, while active (self-powered) tags can have a much longer read range. Typically, the data is sent to a distributed computing system involved in, perhaps, supply chain management or inventory control.
THREATS TO PRIVACY AND CIVIL LIBERTIES
While there are beneficial uses of RFID, some attributes of the technology could be deployed in ways that threaten privacy and civil liberties:
FRAMEWORK OF RFID RIGHTS AND RESPONSIBILITIES
This framework respects businesses' interest in tracking products in the supply chain, but emphasizes individuals' rights to not be tracked within stores and after products are purchased. To mitigate the potential harmful consequences of RFID to individuals and to society, we recommend a three-part framework. First, RFID must undergo a formal technology assessment, and RFID tags should not be affixed to individual consumer products until such assessment takes place. Second, RFID implementation must be guided by Principles of Fair Information Practice. Third, certain uses of RFID should be flatly prohibited.
Technology assessment. RFID must be subject to a formal technology assessment process, sponsored by a neutral entity, perhaps similar to the model established by the now defunct Congressional Office of Technology Assessment. The process must be multi-disciplinary, involving all stakeholders, including consumers.
Principles of Fair Information Practice. RFID technology and its implementation must be guided by strong principles of fair information practices (FIPs). The eight-part Privacy Guidelines of the Organisation for Economic Co-operation and Development (OECD) provides a useful model (www.oecd.org). We agree that the following minimum guidelines, based in part on these principles, must be adhered to while the larger assessment of RFID's societal implications takes place:
RFID Practices that Should be Flatly Prohibited:
ACCEPTABLE USES OF RFID
We have identified several examples of "acceptable" uses of RFID in which consumer-citizens are not subjected to "live" RFID tags and their attendant risks.
We are requesting manufacturers and retailers to agree to a voluntary moratorium on the item-level RFID tagging of consumer items until a formal technology assessment process involving all stakeholders, including consumers, can take place. Further, the development of this technology must be guided by a strong set of Principles of Fair Information Practice, ensuring that meaningful consumer control is built into the implementation of RFID. Finally, some uses of RFID technology are inappropriate in a free society, and should be flatly prohibited. Society should not wait for a crisis involving RFID before exerting oversight.
Although not examined in this position paper, we must also grapple with the civil liberties implications of governmental adoption of RFID. The Department of Defense has issued an RFID mandate to its suppliers, schools and libraries in the have begun implementing RFID, the EU and the Japanese government have considered the use of RFID in currency, and British law enforcement has expressed an interest in using RFID as an investigative tool. As an open democratic society, we must adopt a strong policy framework based on Principles of Fair Information Practice to guide governmental implementation of RFID.
November 14, 2003
The following technological limitations have been proposed as reasons why consumers should not be concerned about RFID deployment at this time. We address each perceived limitation in turn, and explain why in themselves, these limitations cannot be relied upon as adequate consumer protection from the risks outlined above.
1. Read-range distances are not sufficient to allow for consumer surveillance.
RFID tags have varying read ranges depending on their antenna size, transmission frequency, and whether they are passive or active. Some passive RFID tags have read ranges of less than one inch. Other RFID tags can be read at distances of 20 feet or more. Active RFID tags theoretically have very long ranges. Currently, most RFID tags envisioned for consumer products are passive with read ranges of under 5 feet.
Contrary to some assertions, tags with shorter read ranges are not necessarily less effective for tracking human beings or items associated with them. In fact, in some cases a shorter read range can be more powerful. For example, if there were an interest in tracking individuals through their shoes as they come within range of a floor reader, a two-inch read range would be preferable to a two-foot read range. Such a short range would help minimize interference with other tags in the vicinity, and help assure the capture of only the pertinent tag positioned directly on the reader.
2. Reader devices not prevalent enough to enable seamless human tracking.
The developers of RFID technology envision a world where RFID readers form a "pervasive global network" It does not take a ubiquitous reader network to track objects or the people associated with them. For example, automobiles traveling up and down Interstate 95 can be tracked without placing RFID readers every few feet. They need only be positioned at the entrance and exit ramps. Similarly, to track an individual's whereabouts in a given town, it is not necessary to position a reader device every ten feet in that town, as long as readers are present at strategic locations such as building entrances.
3. Limited information contained on tags.
Some RFID proponents defend the technology by pointing out that the tags associated with most consumer products will contain only a serial number. However, the number can actually be used as a reference number that corresponds to information contained on one or more Internet-connected databases. This means that the data associated with that number is theoretically unlimited, and can be augmented as new information is collected.
For example, when a consumer purchases a product with an EPC-compliant RFID tag, information about the consumer who purchased it could be added to the database automatically. Additional information could be logged in the file as the consumer goes about her business: "Entered the Atlanta courthouse at 12:32 PM," "At Mobil Gas Station at 2:14 PM," etc. Such data could be accessed by anyone with access to such a database, whether authorized or not.
4. Passive tags cannot be tracked by satellite.
The passive RFID tags envisioned for most consumer products do not have their own power, meaning they must be activated and queried by nearby reader devices. Thus, by themselves, passive tags do not have the ability to communicate via satellites.
However, the information contained on passive RFID tags could be picked up by ambient reader devices which in turn transmit their presence and location to satellites. Such technology has already been used to track the real-time location of products being shipped on moving vehicles through the North American supply chain.
In addition, active RFID tags with their own power source can be enabled
with direct satellite transmitting capability. At the present time such tags
are far too expensive to be used on most consumer products, but this use
is not inconceivable as technology advances and prices fall.
5. High cost of tags make them prohibitive for wide-scale deployment.
RFID developers point to the "high cost" of RFID tags as a way to assuage consumer fears about the power of such tags. However, as technology improves and prices fall, we predict that more and more consumer products will carry tags and that those tags will become smaller and more sophisticated. We predict that the trend will follow the trends of other technical products like computers and calculators.
RFID Position Paper
A Critique of Proposed Industry Solutions
The RFID industry has suggested a variety of solutions to address the dangers
posed by RFID tagging of consumer products. Among them are killing the tags
at point of sale, the use of "blocker tags," and the "closed
system." We examine each strategy in turn.
KILLING TAGS AT POINT OF SALE
Some have proposed that the RFID tag problem could be solved by killing the tags at the point of sale, rendering them inoperable. There are several reasons why we do not believe this approach alone and without other protections will adequately protect consumer privacy:
Killing tags after purchase does not address in-store
tracking of consumers.
To date, nearly all consumer privacy invasion associated with RFID tagging of consumer products has occurred within the retail environment, long before consumers reached the checkout counter where chips could be killed. Examples include:
We recognize the need for stores to control shoplifting and make general assessments to enhance operations. However, monitoring and recording the detailed behaviors of consumers without their consent, even if only within the store, violates Principles of Fair Information Practice.
Tags can appear to be "killed" when they are really "asleep" and can be reactivated
Some RFID tags have a "dormant" or "sleep" state that could be set, making it appear to the average consumer that the tag had been killed. It would be possible for retailers and others to claim to have killed a tag when in reality they had simply rendered it dormant. It would be possible to later reactivate and read such a "dormant" tag.
The tag killing option could be easily halted by government directive.
It would take very little for a security threat or a change in governmental
policies to remove the kill-tag option. If RFID tags are allowed to become
ubiquitous in consumer products, removing the kill option could enable the
instant creation of a surveillance society.
Retailers might offer incentives or disincentives to consumers to encourage them to leave tags active.
Consumers wishing to kill tags could be required to perform
additional steps or undergo burdensome procedures, such as waiting in line
for a "killer kiosk" and then being required to kill
the tags themselves. Consumers who choose to kill the tags might not enjoy
the same discounts or benefits as other consumers, or might not be allowed
the same return policies. In many areas of privacy law, this retailer incentive
is recognized, and there are legislative prohibitions against inducing the
consumer to waive their privacy rights.
The creation of two classes of consumers.
If killing tags requires conscious effort on the part
of consumers, many will fail to do so out of fear, ignorance, or lack of
time. Many will choose not to kill the tags if doing so is inconvenient.
(The current "killer kiosk" requires loading one item at a time, a lengthy
and time consuming process.) This would create two classes of consumers:
those who "care enough" to kill the RFID tags in their products and those
who don't. Being a member of either class could have negative ramifications.
RFID blocker tags are electronic devices that should theoretically disrupt the transmission of all or select information contained on RFID tags. The proposed blocker tag might be embedded in a shopping bag, purse, or watch that is carried or worn near tags with information consumers want blocked.
According to our understanding, the blocker tag does not yet exist. Until a blocker tag is built and tested, there is no way to know how effective it will be and whether it can be technically defeated.
The blocker tag might encourage the proliferation of RFID devices by giving consumers a false sense of security. While the proposed invention is an ingenious idea, it's one that could be banned or be underutilized if consumers become complacent. It's also possible that such an electronic device could be technically defeated either purposefully or because it stops functioning naturally.
The blocker tag could be banned by government directive or store policy.
Consumers could lose the right to use blocker tag devices if the government deems that knowing what people are wearing or carrying is necessary for national security. They might disallow the devices altogether or name selective spaces in which blocker tags would be disallowed. It is not inconceivable to imagine a ban on such devices in airports or public buildings, for example.
Retail stores might ban blocker tags if they believe the tags might be used to circumvent security measures or if they believe knowing details about consumers is valuable in their marketing efforts.
Once RFID tags and readers are ubiquitous in the environment, a full or partial ban on a privacy device like the blocker tag would leave consumers exposed and vulnerable to privacy invasion.
A blocker tag shifts the burden of protecting privacy away from the manufacturers and retailers and places it on the shoulders of consumers. In addition, busy consumers might forget to carry blocker devices or forget to implement them, especially if additional steps are required to make them effective.
Fails to protect consumers once products are separated from the blocker tag.
Blocker tags theoretically work only when they are close to the items they are designed to "conceal" from RFID reader devices. Once items are out of the range of the blocking device, consumers would be exposed and vulnerable to privacy invasion. For example, a consumer might buy a sweater and feel that the information on the embedded RFID tag is unexposed because she is carrying it home in a bag impregnated with a blocker device. However, once she removes that sweater from the bag and wears it in range of a reader device, information from that tag could be gleaned.
The creation of two classes of consumers.
Like the kill tag feature, blocker tags will also likely create two classes of consumers, those who block tags and those who do not.
Industry proponents argue that when RFID applications
are confined to closed systems, the data is only accessible to those within
the system and those with a government mandate (perhaps via legislation such
as the Communications Access to Law Enforcement Act (CALEA)). Therefore they
argue, society-wide profiling and tracking are not likely. An example of
a current closed application is RFID in libraries. The Grapes of Wrath in
Library X has a different code than the same book in Library Y.
Whereas today RFID applications are confined to closed systems, there will be great incentives to standardize product level tagging. Publishers, for example, may someday ship books to libraries and bookstores with writable tags. Each copy of The Grapes of Wrath will contain a portion of its EPC code that is the same as every other copy. The library will be able to customize the remainder of the code to suit its own inventory control purposes.
Even if closed systems remain closed, their lack of transparency makes them troubling from a privacy perspective. Because details about closed systems might not be readily available, consumers could have difficulty obtaining the information necessary to assess privacy risks and protect themselves.
We appreciate that industry proponents are making an effort to address consumer privacy and civil liberties concerns associated with RFID technology. However, while we believe the proposed solutions are offered in the proper spirit, they provide inadequate protection. Until appropriate solutions are developed and agreed upon, we believe it is improper to subject consumers to the dangers of RFID technology through item-level consumer product tagging.