The Next Big Thing in privacy is the advent of usage-based insurance (“UBI”), made possible by a telematics device – a small gizmo that plugs in to the diagnostic port of a car, monitors a driver’s driving habits, and sends that information wireless to an insurer/third party. Insurers in turn offer up to 25% savings on insurance rates based on “safer” driving. Available for over a year in the US, insurers have now begun offering similar programs in Canada, prompting the industry regulator in Ontario, the Financial Services Commission of Canada (“FSCO”), to release a bulletin designed to outline what it views as acceptable and unacceptable practices with respect to privacy, pricing and marketing of such programs. 
In one Canadian insurer’s program, which is expected to be typical, the device records information relates to vehicle location (“GPS”), time, device connections/re-connections and vehicle identification number (“VIN”).The information is transmitted wirelessly to a third party data service provider, and then erased from the device. Other information, such as kilometres driven, rates of acceleration and braking, speed, cornering and lane changes is derived from the above recorded information. This collected data is then sent from the third party data service provider to insurer, which uses it to calculate the appropriate level of telematics discount, based on distance driven, fast acceleration and hard braking, and time of use.
Is it Personal Information?
The Bulletin notes that while FSCO does not oversee privacy legislation, it is of the view that it is essential that all insurers using UBI programs comply with the Personal Information Protection and Electronic Documents Act (“PIPEDA“) in their handling of personal information. Insurers must demonstrate that they have a procedure for appropriately managing any personal information, and the customer must be informed, prior to enrollment, about what personal information is being collected, who may use or have access to the information, and how it may be used or disclosed, along with what rights the consumer has with respect to the information.
One of the key issues with the current generation of telematics devices is that they do not identify who is driving a vehicle at any particular time – rates are based on the aggregate usage data for the vehicle, not the individual. Despite this, FSCO has taken the position that telematics data should be treated as personal information, even if it is not about the driving behaviour of any one identifiable individual, and should be handled accordingly.
The current business model uses a computer dashboard, wherein the driver can login and view the data collected, much like consumers can now log into their household utility account and see their electrical or gas usage. Unlike utility bills, however, the information for vehicle usage can be traced to an identifiable individual (e.g. who was using the car at what time and where). The privacy implications of such shared accounts have been litigated before: a Canadian telecom provider recently found itself on the wrong end of a $600,000 lawsuit when it produced one global household bill that captured all services based out of that home – including the wife’s personal cell phone. This usage suddenly appeared on the new global bill, and revealed to her astute husband that his (now ex-) wife was having an affair. It’s not difficult to see how vehicle information could lead to similarly unhappy results.
The Bulletin also requires insurers to guarantee that any third party providers who collect or use personal information provide a level of protection that is comparable to that expected of the insurer, and identifies some key points that must be addressed in written contracts with such third parties.
As well, insurers must ensure that no further data is received or accessed after a consumer terminates his or her participation in the program. The FSCO Bulletin is silent whether existing data may be retained; presumably, this will be within the purview of the Privacy Commissioner. Existing programs in the US are clear that the insurer “will retain information collected or derived from the device indefinitely”. Canadian programs are vague, and appear to contemplate that the insurer (the FSCO-regulated entity) won’t keep the data or personal information, but that the (unregulated) third party service provider might.
What Can/Can’t the Information be used for?
The Bulletin also addresses consumer rights, relations and information and outlines several important practices. Specifically, FSCO has indicated that such programs must be voluntary and, at this point in time, a consumer cannot be forced to use a device if he or she does not wish to.
When a user does enroll, all benefits, conditions or exclusions must be clearly articulated in the terms and conditions and enrollment documentation. In the case of amendments to the insurance agreement and other paperwork associated with the insurance policy, the approval of the Superintendent will be required. Insurers may also be required to provide the Superintendent with any advertising or promotional materials, to ensure that promotion of the program is clear, fair and not misleading.
A key point in the FSCO bulletin is that the rating model for a UBI program be clearly communicated to the consumer at all times, beginning prior to enrollment and continuing through each policy renewal. This would include a clear explanation of how to qualify for a discount, the period being measured to calculate the discount, the maximum or minimum of that discount, and regular feedback on driving habits to the consumer. This data should also be made accessible to consumers, who should also be given the opportunity to correct any inaccuracies.
The Bulletin notes that drivers should be able to enroll in a UBI program without also being required to share their personal information for unrelated purposes, including marketing, offering additional services, etc. In other words, drivers should be allowed to opt-in (not opt-out) of non-UBI services. Many of these services would be provided by third party providers of data services, and services that have been proposed include family surveillance, geo-fencing (an automatic alert generated when a vehicle leaves a defined area), or emergency assistance. Some insurers have specifically pitched these additional services to parents as a way to monitor teen’s driving habits, location, and curfews; a US insurer includes an on-board video camera that captures the offending behaviour, is reviewed by a personal driving coach, and issues a weekly report card on teen driving behaviour to parents.
While insurers and data service providers view the strength of these programs as providing incentives to modify behaviour in a positive way (e.g. safer driving, less frequent driving), they will want to be cautious about what behaviours are actually being incentivized. A British insurer recently ran in to trouble with a teenaged insured whose policy allowed him to drive between 5 am and 11 pm. A breach of the “curfew” incurred a £100 penalty – but a penalty for speeding would not be triggered unless he was 50 per cent over the limit. He elected to speed home one evening in order, some say, to meet his curfew. He lost control of the vehicle and he and his passenger were killed.
The foregoing example raises an interesting question – if insurers and third party providers of data services have access to this driving information, and have the capacity to analyze it (either individually or aggregate), does it raise a duty to warn? For instance, if the data collected in aggregate show that young drivers with a curfew are racing home and getting in accidents, is there a duty to warn these drivers? An obligation to modify the metrics? Could this give rise to a product liability claim? Similarly, it the data show that an individual routinely employs unsafe driving techniques (e.g. hard braking, a result of following too closely), does the insurer have a duty to warn that particular insured? The more important question for insurers and third party providers of data services is whether there are ways (e.g. business model, data handling practices, and contractual language) by which insurers can avoid assuming this liability?
The Bulletin also cautions insurers that they must have a process in place to “manage inaccurate or selective data or data reporting”. This includes both equipment malfunction and consumer tampering (for example, unplugging the recording device). The Bulletin doesn’t contemplate malicious hacking, but this is well within the realm of the possible. Presumably, this would all be addressed by the Bulletin’s requirement of appropriate data security and the consumer right to correct inaccurate data.
Who Owns the Information?
FSCO did not take a position on who owns the data generated and recorded by the telematics device. Instead, it required portability, saying “[i]nsurers should, where possible, facilitate drivers using their personal UBI data for the purposes of entering into a contract with another insurer including enrolling in another insurer’s UBI program.” This goal can be achieved with consumer ownership of the data, but is more likely to be met through a licensing/use arrangement wherein the insurer or the third party provider of data services owns the data and licenses it or permits it to be copied.
The current Canadian offering’s terms and conditions are silent on data ownership, and focus on use instead; contrast this with the contractual treatment of the device itself, in which ownership is clearly ascribed to the third party provider of data services that is also the provider of the device.
In the US, numerous States have passed legislation setting out who owns the information collected by manufacturer-installed on-board electronic data recorders (“EDR”) and the circumstances in which it can be used (generally speaking, the information contained in such manufacturer-installed EDRs can be accessed with the consent of the registered owner of the vehicle – not necessarily always the owner, or generator, of the information – or by court order).
Ontario (and the rest of Canada) has been entirely silent on the matter. This does not appear to have thwarted personal injury lawyers or law enforcement officials who seem to be able to access such information routinely and use it to assign civil or criminal liability. In the civil context, the information has become fair game for lawyers during the pre-trial discovery process; in criminal proceedings, it is usually obtained pursuant to a warrant or court order. In both cases, these are on-board devices and the information is obtained from the device itself. In the insurance telematics context, the information will reside with the insurer or the third party provider of data services. In what promises to be a seismic shift in law enforcement, police may no longer have to get authorization to attach their own device but rather get authorization to get existing data from insurers or third party provider of data services, similar to what is now done to get information from internet service providers. In the current Canadian UBI offering, both the insurer and the third party provider of data services are silent as to whether they will release such information to lawyers or law enforcement, or whether they will do so only if compelled.
The most hotly contested information that is, or can be, collected is location data. In Canada, several recent cases involving GPS units being installed on fleet vehicles triggered a series of privacy complaints. The most recent case involved not a GPS unit per se, but the collection and use of information transmitted by GPS enabled cell phones issued by an employer to its employees. The B.C. Privacy Commissioner ultimately concluded that it’s acceptable for companies to collect the GPS data, but with two important limits: informed consent, and purposive use. These cases also involve the interpretation of B.C.’s privacy statute, which has specific provisions that allow companies to collect personal information for the purposes of managing employer-employee relationships, and may not apply in other jurisdictions that lack similar statutory language. It will be interesting to see how these decisions inform an employer’s decision to adopt or decline telematics-based insurance for its fleet vehicles.
Cost and Pricing
While not prohibiting a charge to consumers for the equipment and cost of the program, FSCO has set the expectation that “insurers should cover all costs of enrolling in a UBI program including the cost of any device installation and any ongoing costs of operation or maintenance”. At the same time FSCO has recognized that insurance rates may reflect the development costs of such program.
Finally, mindful that is a new and evolving area, FSCO has said that “at this point in time”, UBI programs should collect and use data solely for discount setting purposes, and not to decline, cancel or refuse to renew risks or to confirm rating criteria currently used. This means that the spectre of a large class of “uninsurable” drivers will not be raised, at least at this juncture.
FSCO has also cautioned that, at this time, it would also be inappropriate for insurers to use data collected for claims-related purposes. It is not clear what this means, but this suggests that UBI program data can be used for rate-setting, but cannot be used to deny coverage, even where such data indicates the driver was at fault or careless. In an earlier presentation, FSCO representatives had said that any discount must be maintained for the term of the policy even if the data from the devices “indicates otherwise”.
The bulletin is silent on whether such data can be used to trigger an independent investigation that results in a reason to deny coverage (e.g. when the data recorded show a series of short starts and stops prior to a collision, the insurer doesn’t use this data to deny coverage, but the unusual pattern prompts it to obtain footage from a nearby CCTV camera – which shows the driver had a friend standing on the roof of the vehicle, “car-surfing”, and uses that information to deny coverage).
The transitory nature of the language suggest that insurers and third party providers of data services, may press for, and FSCO may agree to, changes on these points in the future. Meanwhile, insurers appear to be free to use information gleaned from the aggregate UBI data to inform their marketing approach and policy offerings. In the current Canadian UBI offering, the applicable terms and conditions appear to contemplate that directly: the third party provider of data services is restricted in what it can do, and “can only aggregate this anonymized data and provide third parties with general information that cannot be linked to a specific vehicle (for example, aggregated traffic volumes).” However, even the use of aggregate anonymized data can create difficulty. It was this kind of data which got GPS manufacturer TomTom in hot water when it was revealed that one of the third parties interested in the data were the police, who were using it set up speed traps to capture the very drivers generating the information. 
The release of the FSCO Bulletin marks a seismic shift in how vehicle insurance will be offered in Canada and is an early attempt to set baseline expectations for insurers contemplating such programs. Insurers, and third party providers of data services, will need to look well beyond this Bulletin to ensure their business model and information management practices do not expose them unnecessarily to risk.