Cookies are small text files that websites send to visitors'
computers. Websites struggle to recognise their users without them.
When someone visits OUT-LAW.COM for the first time, our site
endeavours to send that visitor's computer a cookie. We do this
with some help from Google, which offers a free service called
Google Analytics. The Google Analytics cookie is used to count our
visitors, a measure of success that we find rather helpful. This is
just one way in which we use cookies. It's impossible for us to
argue, though, that that cookie is essential to fulfilling the
visitor's purpose.
Under plans endorsed by the European Commission, the Council of
Ministers and the European Parliament, we would have a choice: stop
using Google Analytics or ask visitors for permission to send that
cookie when they visit. Like an over-enthusiastic greeter, the
latter approach requires us to welcome even casual passers-by with
a "Hi, how are you today?" and an invitation to wear a visitor's
badge.
Most websites use Google Analytics (including the site of the
UK's privacy chief, at ico.gov.uk) or a similar traffic
analysis tool, and that is just one use that sites make of cookies.
We’re all subject to this requirement for prior consent – or so it
seems. The trouble is, we don't know what the law really means.
Nobody does, because the proposed law is ambiguous. (See the
relevant sections or
full text.)
There is a cookie law in Europe today. It comes from the
Privacy and Electronic Communications Directive (11-page
PDF), which says that sites using cookies must give visitors "clear
and comprehensive information" about the purpose of the cookies.
They must also offer visitors "the right to refuse" the use of
cookies. That law was passed in 2002 and it is somewhat ambiguous –
but in a way that allows for pragmatic interpretations.
The 2002 Directive did not say when or how the information had
to be provided. It was implemented in the UK in a set of
Regulations that parroted the Directive's ambiguous language. But
our Information Commissioner, to his credit, took a pragmatic view.
He published guidance
(19-page PDF) that said it was acceptable to display the
information in a privacy policy, asking only that "the policy
should be clearly signposted at least on those pages where a user
may enter a website." Usability survived, in the UK at least.
To comply with today's law is easy. Websites add a 'privacy
policy' link to every page, and that policy explains their uses of
cookies. The right to refuse cookies is dealt with retrospectively:
you'll probably have the cookie by the time you read about it. But
that's okay, the Commissioner tells us, provided the policy guides
users on how they can control and delete the cookies on their
machines.
Since that guidance will vary according to the user's choice of
internet browser, we set up AboutCookies.org, a site to
which any privacy policy can direct its readers. Lots of websites'
privacy policies include a link to AboutCookies.org. You'll find
instructions there for controlling and deleting cookies in various
browsers, like Internet Explorer, Firefox, Safari, Opera and
Chrome. (A short time later, the Interactive Advertising Bureau set
up AllAboutCookies, a
site that serves the same purpose.)
This simple approach to cookie compliance is under threat.
The new law says that cookies can be delivered to a user's
computer only if that user "has given his/her consent, having been
provided with clear and comprehensive information" unless it is
"strictly necessary in order for the provider of an information
society service explicitly requested by the subscriber or user to
provide the service."
So if I'm shopping at Amazon.co.uk and I put a book in my
shopping basket, Amazon can use a cookie to remember which book I
want when I proceed to the checkout. That is a cookie that's
essential to the service I've explicitly requested. But if Amazon
wants to use a cookie for another purpose, e.g. to monitor shopping
basket abandonment, it needs my consent.
This sounds bad, but a recital to the new law sounds like an
escape clause. In any Directive, recitals are listed before the
formal 'Articles'. They provide an introduction to the new law,
sharing the lawmakers' rationale for the provisions that follow.
Curiously, the cookie recital includes a suggestion that conflicts
with the main Article.
The new cookie recital says: "the user's will to accept
processing may be expressed by way of using the appropriate
settings of a browser or other application."
Most browsers have a default setting that allows cookies. Most
people never change that (and many don't know that the
setting exists). So a court might reasonably question how
consent can be implied from a default setting. If no question is
asked, silence does not convey consent.
In fact, the expression of a "will to accept" is as close as the
recital gets to mentioning consent. The recital refers twice to
"the right to refuse" a cookie – yet the Article itself tells us
that users must give consent, which is a different standard and a
higher one.
The recital reads like an afterthought, like an apology for
the over-zealous Article that follows. As such, the combination
makes little sense – and websites are given a headache.
A
press release about the Directive said very little about the
cookie plan. All it said was: "The installation of 'cookies' on
users’ computers would also be subject to consent by the user."
We contacted the European Parliament's press office for further
guidance. A spokesman answered: "there has not yet been, at this
stage, any clear and conclusive interpretation."
"Ultimately, whether we're talking about the existing text of
5(3) or the revised one, only courts can give an authoritative
interpretation – there's a functional separation if you will
between the legislator and the court system," he said. "Guidelines
etc. can also be expected to be issued by the Commission, the EDPS
[the European Data Protection Supervisor, who guides EU
institutions on their own compliance] and national data protection
authorities."
"Only the courts can decide" is a lawyer's second-favourite
back-covering cliché, right after "every case turns on its own
facts." It's completely unhelpful to hear it from lawmakers. How
can judges and regulators know what lawmakers intended if they
express themselves in riddles?
The root of the problem is that this law is probably not aimed
at cookies at all. It is aimed at more sinister things being placed
upon or read from website visitors' computers. In an effort to
remain technology-neutral, the Article fails in its purpose. It
talks of storing "information," not cookies, thereby categorising
harmless cookies and password-stealing Trojans together. That is
unhelpful and we have been left with ambiguous wording.
Fortunately, there is time to fix this legislative mess. The law
is part of a wider telecoms reform package and MEPs took exception
to another detail of that package last week, an attempt to combat
file-sharing with a controversial 'three strikes' rule. They
sent the whole thing back to the
European Commission as a consequence.
If the Commission accepts the file-sharing changes, it will go
to the Council of Ministers on 12th June. The Ministers will decide
whether or not to accept the Parliament's changes and any further
changes by the Commission. By then there will be a new Parliament
and, from November, a new group of Commissioners. Despite the fact
that all three bodies had agreed the cookie part of the law, the
new assemblies may revisit and change that bit too.
To anyone seeking imminent election to the European Parliament:
will you fix this please? It's not the first time that cookie plans
have needed fixing: the same problem arose with the original
Directive (see: European Council votes for
spam opt-in and new cookie plan, OUT-LAW News, 07/12/2001), but
the end result was acceptable. For companies with EU
lobbyists: for the sake of your website's usability, please ask
them to lobby.
Had MEPs not taken exception to the file-sharing provisions, the
latest cookie proposals would have been installed in EU law this
month with the stealth of the spyware they set out to block. Most
of us did not know that this was going on: the English text with
the cookie provisions became available only on Thursday. But our
silence must not be misinterpreted as consent.
By Struan
Robertson, editor of OUT-LAW.COM. This article
represents Struan's views – not necessarily those of Pinsent
Masons. Follow Struan's 140-character rants and ramblings at
twitter.com/struan99.
Update, 21/05/2009
The ICO has responded to our request for comment with the
following statement:
“The ICO agrees that it would be impractical
if a narrow interpretation of any revision to the Directive
required users having to consent to the use of cookies. The
ICO’s view has always been that users’ interests and privacy are
sufficiently protected by the provision of information about
cookies and the opportunity to refuse them.
“As the legislative process is ongoing there
will be opportunities for all stakeholders to influence the
debate.”
A spokesperson confirmed that the ICO will monitor progress of
the law in Europe. "We're not going to make a call on it until
the process restarts,"
she said. "If we were to raise a concern, it would be through the usual channels:
the Article 29 Working Party, the EDPS [European
Data Protection Supervisor] and BERR [Department of Business
Enterprise and Regulatory Reform]."
