Privacy Law and Policy Reporter
Robin Whittle conducts a dialogue with Telstra.In late 1991 Telstra (then Telecom) first signalled to AUSTEL its desire to introduce Calling Number Display (CND) -- or Caller ID as it is known in the US. This, together with the problem of telemarketing, led to AUSTEL's Privacy Inquiry which concluded late in 1992, recommending that any introduction of CND be based on the principle of informed choice. Three years later, AUSTEL's Privacy Advisory Committee (PAC) has released its report on the contentious service (Calling Number Display, AUSTEL, January 1996). The 100-page CND report is available free from AUSTEL (+61 3 9828 7300), or via the World Wide Web at http://www.austel.gov.au
This article examines key points of the PAC report, describes progress on technical standards for CND and reports discussions of Telstra's CND plans with Group Manager of Corporate Policy, Michael Pickering.
The PAC had requested specific proposals for CND introduction from Telstra and Optus, but these were not forthcoming and it began work in July 1995 on the basis of opt-out, with free per-line and per-call `blocking' and free per-call sending of the caller's number. The report was released in January 1996, recommending that CND be introduced on the `opt-out' model, in which callers' numbers will be displayed to the people they call unless they take special action to prevent it. While it deals primarily with the display of the caller's number, and with the use of that number for returning calls, it also mentions the display of the caller's name -- without exploring the privacy and administrative problems such a service would raise. CND and Call Return involve no extra overhead in the existing inter-exchange signalling, but a name display requires additional transactions to find the name from a remote database.
Each of these is discussed in detail in the second part of this article in the next issue of PLPR.
He said that `AUSTEL's PAC report indicates a starting point for a Calling Number Display service. Now we can develop models of demand and costs, including the cost of the public education campaign, to build a business case for the service.' He noted that Telstra would only proceed if their modelling indicated the service would be financially viable in the long term and that at the earliest it would take 12 months before introduction could begin.
Telstra's `Future Mode of Operation' upgrade of its telephone exchanges will mean that the CND display service could technically be provided in most metropolitan areas by the end of 1996. WG 2/1's work is directed at providing an electrical and message format standard for customer equipment (display boxes, special phones or modems) to receive the CND information from the exchange. On the analogue lines used by all residential customers and many small businesses, including those with `Commander' systems, this is achieved with three or so seconds of 1200 baud modem tones -- between an initial short burst of ring and the standard ringing sequence -- which must be delayed by three or four seconds. There is no provision initially for the exchange sending the number of in incoming call to the display device while a previous call is on hold as part of a call-waiting service.
The message could comprise a caller's number and or name, with flags to indicate why each is not available. These may be `Private' -- indicating that the caller chose, on a per-line or per-call basis not to allow the display of their number, or `Unavailable' indicating that the call came from a service, or via a network, which was not capable of providing a number for CND display.
WG 2/1 has no connection with the PAC. It works under instructions from AUSTEL's Standards Advisory Committee -- which requires that the technical standard not specify exactly how the number, name, `reason for absence' flags or any other kind of message should be displayed. There would be consumer advantages in standardising the displayed text -- especially the `private' and `unavailable' messages, for which alternatives such as `number blocked' or `out of area' are sometimes suggested. While the Standards Advisory Committee wants the form of the display to be a matter for the market (manufacturers and consumers) to choose, the PAC report states that:
The terminology used for these messages ... should also be consistent across technologies and networks in the interests of ease of understanding and use for consumers, and to enhance the call management advantages of CND services.Unfortunately the PAC suggests using the term `blocked' -- which carries negative connotations -- instead of `private'.
Other kinds of analogue lines, such as those which interface to analogue PABXs are not covered by this standard. It seems unlikely that the majority of PABXs which use decadic or DTMF signalling will ever be capable of receiving CND. However many large PABXs, and in the future more and more smaller ones, connect to the network with ISDN (Integrated Services Digital Network) lines -- which fully support CND messages.
CND display is part of the GSM standard, so digital mobile phones are already capable of displaying CND, but analogue mobiles are incapable of receiving the CND messages and so will never display the caller's numbers.
For the caller to enable or disable the display of their number on a call-by-call basis, two prefixes would be required and the PAC report states that these will be allocated from the 183x number range -- or perhaps #83x or #3x on tone-dial and mobile phones.
After agreeing that the highly publicised COT cases of past years (see 1 PLPR 52, 115), experiences of poor customer service and the present inability to detect or deter one-off malicious calls were negative influences on perceptions of the company, we discussed Telstra's view of Calling Number Display.
Telstra wishes to provide a display service which will enhance the privacy and phone management capabilities of those residential customers who subscribe to it -- as well as providing efficiency gains and the possibility of new services for business customers. Pickering describes it as `making a body of existing network data accessible to the users of that network.' Carriers do not envisage a situation in which callers will pay to display their number, so the service is charged for at the receiving end and its value is seen to be dependent on a high proportion of incoming calls having displayable numbers.
AUSTEL's research did not find any instances where overseas carriers had researched the prospect of callers paying to have their number displayed, nor any instances of CND being introduced on an opt-in basis. However, there are high-value, trust-binding and service-enhancing applications of CND for which callers might be happy to pay a modest fee per call. Thus the prospects of additional revenue from callers in an opt-in introduction (which has far fewer privacy problems and public education costs) do not seem to have been researched.
The lifeblood of a CND service is the caller's phone number -- or the number of the service from which the caller is dialling. In the context of particular phone calls, we agreed that this often constitutes personal information. Privacy proponents aim to ensure that the number is displayable only when the data subject -- the caller and/or the customer whose line the call is being made from -- has made an explicit, well-informed choice to allow the display of the number. There is a counter-argument that CND provides privacy protection and other benefits to the receiver, but in either an opt-in or opt-out scenario, CND offers few `hard' privacy benefits -- since callers making unwelcome calls are likely to do so without a displayable number. Nevertheless CND, especially on an opt-out basis, may lead to many `hard' privacy problems for callers, including people who are unaware their number is being displayed.
Privacy advocates generally agree that CND has some benefits, and would accept or welcome an opt-in scenario, while the carrier's imperative is to get regulatory approval for an opt-out introduction with as few impediments as possible to calls having displayable numbers. The debate revolves around questions of what constitutes adequate privacy protection for consumers who make no choice about the default status of their line, about the benefits of the CND service (for callers and receivers) and about its potential for misuse.
The first introduction of Caller ID by New Jersey Bell in 1988 involved `compulsory' display -- most callers had no choice about their number being displayed. The oft-quoted reduction in harassing call complaints in this introduction is not relevant to the opt-out or opt-in models being debated in Australia -- since in these models, nuisance callers would generally make calls without displayable numbers. While supporting Call Trace as a better approach, AUSTEL's PAC is generally upbeat about opt-out CND's ability to reduce malicious calls. The PAC seems unaware that the New Jersey CND introduction coincided with that of Call Trace and `Automatic Callback' -- a combination which led the Public Utilities Commission of Ohio in 1992 to conclude that `it is impossible to determine which, if any, custom-calling services contributed the most to the decline.'
Telstra is currently unable to trace malicious calls (threatening, harassing or offensive as defined in s 85ze of the Crimes Act) if they occur just once to a particular victim. Tracing can only be instigated subject to a written request and Telstra's judgment about the seriousness of the situation -- and so is little deterrent to malicious callers who make any number of isolated calls.
After discussing the company's plans for improving malicious call tracing in the next year or two, Telstra suggested that opt-out CND was a really good approach to stopping malicious calls. I responded that it was a very poor approach since, in order not to answer malicious calls, the customer must refuse to answer all calls without displayable numbers. This would include calls from payphones and may involve putting up with the phone ringing repeatedly -- a form of harassment in itself. Customers seeking to use CND as a defence would need to pay for the service (while Call Trace has no monthly charge), buy equipment (Call Trace needs none), and look at the display box before deciding whether to answer each call.
In conversation, some of these points were conceded, but Telstra's considered response is that my positive assessment of customer-activated Malicious Call Trace is overly optimistic and that its deployment requires greater analysis. Citing the reports of Wauchope residents, Telstra maintains that CND would assist in the management of nuisance calls and that CND customers will be no worse off, and in some circumstances better off with respect to malicious calls. This response does not address the evident difficulties and expense in using CND to protect against malicious calls. The Wauchope evidence is insubstantial and may be biased by the fact that some residents who reported malicious calls were given a CND service -- when in ordinary circumstances they may have been offered the existing arrangements for call trace.
Customer-activated Call Trace has nothing to do with CND -- the victim simply dials a code after receiving a malicious call, so the exchange captures the number of the previous incoming call and forwards it directly to trained investigators. Any malicious caller, including those using payphones, who makes a series of calls will soon come to the notice of investigators, who would have much more solid evidence than a victim's report of their CND display. Some problems with customers activating the trace by dialling immediately after the call can be overcome by another approach involving verbal reports to investigators who access exchange records to subsequently trace the source of the problem call.
There are many problems raised by CND. One simple example is dialling the displayed number of an unanswered call in an attempt to contact the original caller -- which Canadian researchers label a `blind call-back'. Such call-backs may not be welcomed by the original caller, and may occur at inconvenient times or be answered by someone who should not be involved. The PAC report states that `30 per cent of (surveyed) Canadians considered blind call-backs to be an invasion of privacy'.
Although no figures for the level of incidence of such calls is available, the incidence could be quite high since many carriers offer a CND based `last call return' service, and Wauchope respondents reported that the recording of unanswered call's numbers was the most important advantage of CND.
The privacy advocate's position from the caller's perspective is that their number is very often personal information and that CND involves the carrier selling that personal information. This should only be done with the explicit permission of the data subject -- the caller, or the customer whose line the call is made from. Hence any opt-out scenario violates this most basic privacy principle, establishes casual disclosure as a social norm and will lead to widespread problems including abuses by organisational CND customers and unintended breaches of personal information boundaries in family, friendship and business relationships. While the proportion of calls involving CND privacy problems may be low in percentage terms, Telstra reports that more than 35 million calls are made each day.
Many privacy advocates argue that since CND provides limited benefits for consumers, it should be introduced on an opt-in basis (which provides similar benefits) or not at all, and that other means, such as Call Trace be used to deter and detect both one-off and persistent malicious callers. v
The second part of this article, dealing with the PAC's conditions for CND introduction, will be published in the next issue of PLPR. Robin Whittle is a consultant on hardware, software and human interaction in telecommunications and music, and a consumer advocate on telecommunications policy. Contacts: tel Ph +61-3-9459-2889; home page http://www.ozemail.com.au/~firstpr ; email email@example.com
 For brevity the term `caller's' is used, while in fact the number or name relates to the phone service from which the call is made, rather than to the person actually making the call.
 The author represents Consumers' Telecommunications Network on this working group.
 Whittle, `Telstra's New Model Network', Australian Communications, April 1996.
 Chapter II, 4.4.
 Public Utilities Commission of Ohio, Decree on Caller ID, (II, E, 5) 26 March 1992. http://www.eff.org/pub/Privacy/ Caller_ID/Foreign_and_local/OH/ohio_ cnid_puc.order
 Sterley and Associates, Wauchope caller ID trial, pp 88-89: `Very few residential (qualitative survey) respondents had any experience with unwelcome or prank calls. Three respondents had had prank calls in the past'. Business respondents reported that prank calls ceased when it became known they had CND displays. No quantitative data on prank calls is reported. On a national basis, Telstra keeps no records of malicious call complaints.
 Sterley and Associates, Wauchope caller ID trial, p 77: four customers were given CND services because of problems with nuisance calls, but there is no mention of them being offered Telecom's standard call trace. Customer-activated Call Trace was not mentioned or researched.
Chapter II, 3.1.2.
 Sterley and Associates, Wauchope caller ID trial, p 98.