Examination of Witnesses (Questions 100-119)
Professor Elspeth Guild and Mr Tony Bunyan
21 MARCH 2007
Q100 Lord Jopling: Surely, it cannot
be a surprise to anybody that mistakes occur with all these millions
of tickets being sold. What do you say when somebody says to you:
"Well, so what? Mistakes are made and inconvenience is occasionally
caused out of all those millions, but, surely, that is a very
small price to pay for the overall benefit given the terrorist
threat that faces us"? Listening to what have said so far,
if somebody said to you: "Well, all right, we hear what you
say but you are scratching around on the periphery of all this",
what would you say to that? I know I am being aggressive.
Mr Bunyan: It is a question one is often asked,
and if the purpose of PNR is to tackle terrorism, fine, but it
is wider than that. It is wider in terms of organised crime, money
laundering and, in some cases, other crime. What we have got to
do (and this, again, goes into the second subject as well) is
be much more targeted. If it is just to do with terrorism then
the numbers would be much smaller on the watch lists than they
are. In the United States the watch list is 132,000; we do not
know how big our UK watch list is but we know what theirs is,
and that I can understand. However, when you start to say: "While
we are collecting information to stop terrorists coming we are
going to stop criminals (undefined) and illegal immigration people
(undefined)"- it may be a minor offence or a major offence,
we do not know—
Q101 Lord Jopling: Why not?
Mr Bunyan: I suppose you might want to live
in a different kind of society where everybody is checked for
everything. I think in a democracy one has got to actually say
that if the intention is to tackle terrorism let us have measures
to deal with terrorism, and let us have guidelines and accountability
in relation to that. If we want to extend it into crime, let us
have that discussion.
Lord Harrison: Is there not overlap?
My instinct is that if you were to look into this—and I
understand from what you are saying that we have not got the kind
of breakdown which would be very useful to us—I guess in
that penumbra of those who fall into the category of criminal
activity, they may well draw from that number, on occasions, potential
terrorists.
Lord Teverson: Terrorists are probably
the most law-abiding people in society!
Q102 Lord Marlesford: I want to go
back to Professor Guild's 40 per cent. What did that 40 per cent
figure refer to? You may have told us and I missed it.
Professor Guild: There are two categories. These
are from the reports of the Data Protection Supervisors in three
EU Member States on random checks which they have carried out
regarding information in national Schengen Information System
files. In their reports, over the last three years, they have
been reporting that of the controls which they have carried out
on information which they have checked there have been factual
errors, incorrect storage or storage of data which should have
been deleted in up to 40 per cent of the cases which came before
them.
Q103 Lord Marlesford: That 40 per
cent figure does not mean very much unless you split it up.
Professor Guild: The vast majority of the problems
were inaccuracies in the data.
Q104 Lord Marlesford: What you are
saying is that the vast majority of the 40 per cent had an inaccuracy,
or one or more inaccuracies in the data?
Professor Guild: That is my understanding. I
have not read the reports myself.
Q105 Lord Marlesford: Out of 36 questions
(whatever it is) there might have been one inaccurate one in 40
per cent of the cases.
Professor Guild: That could certainly have been
the case, but it could also have been more.
Q106 Lord Marlesford: Or more than
one.
Professor Guild: More, yes.
Q107 Lord Marlesford: In other words,
going back to the Chairman's point, the more questions you ask
the more information you are going to get, both accurate and inaccurate.
On that basis, it seems to make the opposite case to that which
you are making.
Professor Guild: There is another issue which
comes back to the question of risk. When you collect data, when
information is collected by airlines and passed to governments,
how many spelling errors are tolerated before the person is refused
the right to travel by the airline? If you type in your first
name incorrectly, are you refused access to the `plane? If you
type in an extra letter in your surname, does it prevent you travelling?
If your gender is incorrect because you ticked "Mr"
instead of "Miss", are you prevented from travelling?
The information which is passed on is the information which you
typed in online, with all of the human inaccuracy which is inevitable.
If that is the information against which checks are being made,
the question is: is this a system which helps law enforcement
or, in fact, has any effect on risk? If you are intending to blow
up an aeroplane with a suicide bomb, if you have one letter wrong
in your surname and your gender wrong and you still travel, the
information which has been sent is the information which you provided.
Is this an effective way of dealing with the question of risk?
Will that information be associated with the name on the watch
list?
Q108 Lord Marlesford: You are still
making the same point that I was making. If you only ask people
their name and you decided whether or not to let them in on the
basis of their name and they got it wrong, and you decided you
would not let them in on that basis, that would be a much greater
risk to people than asking them 40 questions on which they might
make one or more mistakes.
Professor Guild: Increasing the number of elements,
of course, will increase the risk of error. These checks are being
done automatically by computer programmes, there is no individual
who can say: "That is clearly not an important one, but let's
look at that".
Chairman: The well-publicised case of
Edward Kennedy presumably arose because there was another Edward
Kennedy. The passport numbers and the dates of birth, and so on,
must presumably have been wrong. Anyway, I will ask Lord Foulkes
now, please.
Q109 Lord Foulkes of Cumnock: I travelled
halfway round the world as Mrs Foulkes, but I do not see what
relevance that has to what we are discussing today. What worries
me, my Lord Chairman, is that a number of statements and alleged
facts are being made, particularly by Mr Bunyan, without giving
us corroboration. Who says there are 132,000 on the US watch-list?
Where did you get that figure from?
Mr Bunyan: Most of my information is coming
from the Government Accountability Office of the United States,
which is a highly respected organisation. So respected, in fact,
that Congress shuts down a measure if they come up with a bad
report, as it did over CAPPS II.
Q110 Lord Foulkes of Cumnock: From
whom?
Mr Bunyan: The Government Accountability Office
of the United States.
Q111 Lord Foulkes of Cumnock: As
a government?
Mr Bunyan: It is a bit like our National Audit
Office except it does a lot more, in my view, detailed and important
work. It is very similar to our National Audit Office except it
has far more resources and goes really into depth about the effectiveness
of privacy and data protection issues.
Q112 Lord Foulkes of Cumnock: You
said earlier that when I book a flight from Edinburgh to London
by British Airways I am agreeing to all my information being passed
to the United States authorities.
Mr Bunyan: I have to tick a box or I cannot
book online.
Q113 Lord Foulkes of Cumnock: I book
online regularly, and I have never ticked such a box. On what
basis do you say that?
Mr Bunyan: Maybe you are a frequent flier. All
I know is that that particular page we have actually got on our
website. I am telling you my own experience, and maybe you have
a different experience. I am merely giving information, and we
do have a copy of that page on our website, so I could
send you the URL if you doubt what I am saying. We may just have
had different experiences; that is possible.
Q114 Lord Foulkes of Cumnock: Professor
Guild, the 40 per cent figure that you said was from three studies,
what is the basis for that? I was not clear. Maybe I did not hear
it.
Professor Guild: The national reports of Data
Protection Supervisors over the last three years in France, certain
Lander in Germany and in the Netherlands.
Q115 Lord Foulkes of Cumnock: Where
did you get that figure from?
Professor Guild: It is a study which has been
carried out by a researcher at my university who has examined
...
Q116 Lord Foulkes of Cumnock: Where
did she get the information from?
Professor Guild: The public, annual reports
of the data protection authorities in France, some Lander
in Germany, and the Netherlands.
Q117 Lord Foulkes of Cumnock: You
can let us have a copy of that report?
Professor Guild: Of course.[6]
Q118Baroness D'Souza: The Home Office Minister
told us that the PNR data is a valuable source of data for risk
assessment and intelligence purposes. She mentioned also a pilot
project that is currently being undertaken by the UK e-borders
programme. Could we have your views on that?
Mr Bunyan: As I said earlier, it would be good
for this Committee to know and to ask the Home Office what PNR
data the UK is collecting so this Committee could compare PNR
data being collected by the UK, Canada and the United States.
I think that would be a logical question to ask. The pilot programme
she is talking about is a Semaphore programme which is targeted
at flights in and out of particular countries. One might guess
and expect them to be countries like Sri Lanka and Pakistan, where
they are profiling people to a much greater degree than general
travel. This is what that scheme is, although, on the profiling
concept, if you look at the Home Office plan published in December,
while it is currently being used for targeting flights from what
they call "risk countries" they do intend, in the longer
term, to profile all passengers. They are profiling these people
for what, is known as APIS, Advanced Passenger Information System,
which is the red, yellow, green system, which is operated in Australia
now. Red, you cannot get on a `plane, basically, and green you
can get on a `plane. Again, the studies from the United States,
the Government Accountability Office, show that the yellow category
is the biggest problem area, in that people are misidentified
or wrongly identified. We also have a problem down the line in
the UK and the EU that all the studies show, as does Peter Hustinx's
view as a European Data Protection Supervisor, that the bigger
these databases get the error rate increases. That holds for both
the data that is being checked or indeed for fingerprints themselves,
which are meant to be exclusively biometric. All the evidence
shows that the error rates, in other words being wrongly identified
as being someone of risk, or not being identified, which one could
argue is even worse—I am a risk but they have not found
out I am—increase. So, for all the biometric tools, we have
got to look at it very carefully and draw on experiences from
the United States where they have, in my view, much more detailed
information than we have available in the European Union.
Q119 Baroness D'Souza: Surely biometric
data is not subject to the same rate of error as other data?
Mr Bunyan: No, but, again, all the studies show—and
the first ones were done about the Visa Information System, which
in the European Union is eventually going to grow over ten years
to 70 million (ten fingerprints of 70 million)—that as it
gets bigger the error rates are going to grow. The error rates
on fingerprints will not be as great as the errors on data, or
intelligence, but there are error rates which have to be taken
into consideration. We ought not to think that these are totally
infallible systems which the bigger it gets it is going to work.
I give you another example: at the moment we have what are called
digitised pictures in our passports—they call them "biometric"
but all they are is a digitised copy of the normal passport picture;
they are not a biometric—
6 See supplementary evidence from Professor Guild,
p 39. Back
| 
