Examination of Witnesses (Questions 440-459)
Professor Angela Sasse, Professor Martyn Thomas and
Dr Ian Forbes
27 FEBRUARY 2008
Q440 Lord Peston: I understand your
point about over gathering, but where you said "individuals'
entitlements", I thought you were talking about public services
and what I am entitled to because of my specific condition; about
my specific condition; whether I am disabled or a single parent
or this, that or the other means knowing very precisely who I
am.
Professor Thomas: No, it does not actually.
It means knowing what your specific condition is.
Q441 Lord Peston: Therefore a
fortiori I would have thought your argument would be the reverse:
you need to know that detail but then you need absolutely to protect
it from anybody else getting near to it.
Professor Thomas: Actually very rarely
do you need to know who somebody is. If you have to have a proof
of age card that shows you are entitled to buy age-protected goods,
why does that need to have your name on it? There is no reason
why you should communicate to somebody selling cigarettes or alcohol
or letting you into a film what your name is because that is completely
unnecessary.
Q442 Lord Peston: What about my travel
pass? You definitely need a proof of age to get your travel pass.
Professor Thomas: You need a proof of
age to get the travel pass. Once that has been established, it
does not need to say who you are.
Q443 Lord Peston: So everybody could
use the card.
Professor Thomas: No, you can have a
photograph on it. It needs something to link it to you but it
does not need to be your name.
Q444 Lord Peston: With my photo almost
anyone could get by.
Professor Thomas: Having your name on
it is not going to stop somebody else using it because your name
is not tattooed on your forehead.
Q445 Lord Peston: What I am really
trying to say is that we are making a bit of a song and dance
about this, when the real issue lies elsewhere. To go back to
your example, you said you do not want anybody to know about your
Oyster card, so you pay cash. Now that is up to you, but it does
seem to me to be slightly ridiculous, if I may say so, though
it is your choice.
Professor Thomas: No, no.
Q446 Lord Peston: You are entitled
to make that choice and I am not trying to stop you. I thought
we were talking about people entitled to things from the public
sector where it is vital we get the right person.
Professor Thomas: There is a fundamental
point here. Most people, for most of their lives, do not need
to conceal much about themselves but some people, and probably
most people at some time in their lives, need to conceal something.
If everybody in general gives information away, it makes it very
hard for the people who have an entirely legitimate reason—they
are trying to escape an abusive relationship and they do not want
the details of where they are currently living to be known—and
you need to give those people the freedom to behave in a way that
does not immediately highlight them as somebody with a problem.
Q447 Lord Smith of Clifton: The CRC
says that the same technology is capable of affecting different
individuals or different groups in very different ways and it
underscores the Royal Academy of Engineering's concerns over the
inequality by pointing out " ... there will be some individuals
and groups who benefit and some who are harmed". Can you
give some concrete examples of these differential effects? Do
they raise human rights issues? How might these problems be addressed?
Dr Forbes: There is a range of examples
but not very many. One is the way that some call centres know
who you are when you call and they then check you against their
database, whether you are a big spending customer or not. If you
are, you will get through immediately whereas everybody else gets
shoved down the line and they are waiting for 15 minutes, except
if you are waiting 15 minutes, you do not know that is what has
happened, you do not know that is their policy and it is a clear
discrimination against people on the basis of their spending power.
That might be okay if you are informed, but you are not. Then,
by inference, we know that discrimination occurs in very predictable
ways across society within organisations. There is institutional
discrimination and we are talking here about organisations which
do not have to reveal what they are doing, why they are doing
it, how they are doing it. They do not have to assure us that
they are not being indirectly discriminatory, that is to say not
meaning to discriminate against ethnic minorities or economically
disadvantaged people, but that is in effect what they are doing
without justification. We are talking about dealing with huge
groups of people, so we could expect that some systematic disadvantage
is going to be introduced just because people have not been trained
to do otherwise. They are not trained to deal with data, they
are not trained to deal with understanding that all organisations
need to implement the social values to which we subscribe and
for which there is specific law in relation to the provision of
public services and private services. The predictable or usual
suspects get disadvantaged here. For example, we have already
had the evidence about the number of black people on the DNA database.
What is that all about? That seems to me to exemplify discriminatory
assumptions about a group in society; that they are more criminally
active and more likely to commit crime.
Professor Thomas: And reinforces those
views.
Dr Forbes: And reinforces those views.
We know that it is in fact not the case. I guess we know since
the big Sex and Race Discrimination Acts of 1975 and 1976 that
unless you take active steps to reduce discriminatory behaviour
it continues. In this area there are not, as far as I can see,
active steps being taken by any of the organisations or required
in any of the legislation to address these issues. Software programmes
are being written which embed assumptions and stereotypes. The
classic example of course goes way back to St George's Medical
School which used to pride itself on the range and the ethnic
diversity of its input until somebody looked at the programme
used to select and it noticed that they offered places to the
people with lower scores. If you were a woman, you got an extra
10 points. If you were from an ethnic minority you got an extra
20 points. So anybody filling out these forms was unintentionally,
unknowingly, leading to a discriminatory output. Unless we know
how these programmes are written and unless they are proofed in
the appropriate way, then we can actually predict that they will
discriminate. In my view direct action needs to be taken.
Q448 Lord Lyell of Markyate: Why
do you say that was unintentional?
Dr Forbes: The person filling in the
form gets the form, they fill in the age, the gender, the ethnic
origin and behind that, the computer assigns value to it. It produces
a printout, top of the list lots of white males and some white
females and some very, very bright Asian candidates and they are
the ones offered the places. The person doing the data entry had
no idea that the programme writer put those values on at some
point in the past.
Q449 Lord Lyell of Markyate: Yes,
but St George's meant it; they were intentional. That is exactly
why they did it.
Dr Forbes: Yes, the person who wrote
it intended to do that and it may have been legal when it was
written. This was in 1980. It may have been legal before 1976
and 1975.
Q450 Lord Rodgers of Quarry Bank:
Because the audiology technology of the House of Lords is sometimes
defective I may have missed some of your replies so forgive me.
Going back to the RAE report, it recommended a digital charter
which would have a significant effect on the levels of trust.
What effect do decreasing levels of trust have upon the democratic
governability of the country? Would a digital charter add any
real value to the policies, laws and other forms of protection
which we already have?
Dr Forbes: On the question of trust,
this is even more complex than CCTV because it cuts in so many
different directions. I do not believe there is a general decline
in public trust in Government but there are specific instances
in which trust goes up and down. There are so many problems here.
One is that there is misplaced trust in Government to do certain
things that they cannot in fact carry out. Another is a misplaced
distrust which is even tougher for a Government to handle because
it is doing the best it can but it does not get credit for it.
There is evidence that people do not believe that big data can
be secured and yet we have governments continually introducing
measures to deal with big data and promising that it will be secure.
That is a real problem for Government. It needs instead to be
more accurate and say "We can protect this up to a point.
We can put these measures in which are going to reduce the functionality
in some respects or reduce the risk in other respects and we need
to talk about that". A digital charter, if it were to set
out some clear bases for operation, some clear guidelines which
the population and the Government could talk about and agree on
where they are setting the balance, would be tremendously helpful
for encouraging trust in this area because the Government rely
on people to give accurate information. If you are the person
who says actually you are not going to give information because
you do not trust this, then that can do tremendous damage to the
whole exercise of gathering the data. So it is quite important
that we have trust in these systems and that we have trust in
the exercise itself and the purposes for which it is being developed.
We see a lot of examples of CCTV being used for public benefit
but we see very little data which shows how data is effective
in making good government, in allocating resources effectively
or efficiently. A digital charter needs to look at all those issues
so that in a consultative basis it allows people to say what it
is that they want from these systems, from these digital systems;
what it is that they expect. We know that in terms of trust people
mostly think in terms of risk; how risky it is if they give you
this information. So they will trust you if they think the risk
is appropriate. They do not know what the risks are most of the
time; they do not know what the expectation is. It is very difficult
to know, so that is why I think a charter in advance is a way
of getting people to come to a settlement for the time-being,
to start a process which then can be reviewed in the future and
evaluated to see exactly how well we are doing here. Otherwise
it is just stumbling along, we do not have any guidelines and
we do not have any sense of placing ourselves. It seems to me
we need these things quite urgently because of the fast proliferation
of technological mechanisms and means which take data out of our
control and do things with it like data-mining and cross-referencing
which we do not know is happening behind the scenes but which
can have a huge impact on us as citizens.
Professor Sasse: May I add something
to this which is an issue we have not raised so far and that is
data quality? There is often an assumption when data is collected
that it is all correct and it is all used in the right way. If
you look at the reports by the Office for National Statistics,
when you look at records you find that up to 40 per cent of records
are either out of date or contain at least one significant false
bit of information. To me that is something that should be enshrined
in a digital charter, that citizens have the right to check the
information Government hold about them and that it is corrected
if it is not accurate. If decisions are being made about people
that is one thing, but if they are made on inaccurate information
that is another. We have had submissions from people who have
said that their health records, for instance, incorrectly stated
certain things and that they either had a very long battle to
have that corrected—and that again is only something that
the knowledgeable and the wealthy can afford to do—or even
that they were told that there was no way that a record could
be corrected if an entry was older than 90 days and therefore
the best they could do was have post-its with "This patient
is not an alcoholic" plastered all over the hospital to deal
with the fact that the record was incorrect. That strikes me as
an example of the duty of care needing to be that the records
held are correct, that they can be inspected and they can be corrected.
Q451 Baroness O'Cathain: We have
the right, have we not, to ask about our credit rating?
Professor Sasse: Yes, you have the right
to see your credit record.
Q452 Baroness O'Cathain: So really
what you are saying is that it is only Government that withholds
the right to see the records they hold on us. Do we have the right
to see records held by anybody else on us?
Professor Thomas: Yes, under the Data
Protection Act.
Q453 Baroness O'Cathain: But the
Data Protection Act does not extend to the government information
being held on us.
Professor Thomas: It extends to a large
part of Government. Unfortunately, you do not know who holds all
the data so you do not know whom to ask.
Q454 Baroness O'Cathain: I am talking
about NHS data.
Professor Thomas: Secondly, the Data
Protection Act allows the organisation holding the data to charge
you £10 for every database that has to be interrogated and
that is just an impossible barrier to getting access to the data.
You need a clear statement of who has data on you and then you
should have free access to it.
Q455 Lord Peston: I do not want to
appear an anarchist, but are we not right not to trust the Government?
Is that not the nature of democracy? To take the example when
Viscount Bledisloe asked you about medical records, if you were
to ask the public at large "Who do you trust, the Government
or your doctor?" we know exactly what the answer is going
to be, which is what worries one. I am a strong believer in access
by the medical profession to my medical records but what I do
not like is the discovery, and I am told we were doing this and
I did not know, that the Government are proposing to let all sorts
of other people look at my medical records. I do not want a charter,
I just want that not to happen under any circumstances and that
is why, having a charter, okay, is a start maybe to give you access
to check the accuracy. What we want is a basic stop, stopping
the Government saying "Ah, that is a good idea, let's add
them and them and them and add this additional data" and
so on. Is that not what we really need to do?
Dr Forbes: One of the things that the
Academy recommends is that we differentiate between the state
and the Government. We allow the state to gather information about
us, but we set up non-governmental authorities to hold that data
so that the Government, if they want that data which is ours,
has to apply and that makes it a public act, an accountable act
and it is free to apply and ask for whatever it wants for whatever
nefarious purpose, but then we get to know. That is the crucial
thing. That is why I would use a charter: laying down these quite
strict things within which everybody has to work and we understand
what they are. As for trusting Government, most people are quite
happy to let Government get on and do it; they do not really want
to be bothered. That is a form of trust.
Q456 Lord Peston: Yes, but it is
a negative form, is it not?
Dr Forbes: Compared with what?
Q457 Lord Peston: Compared with asking
them the straight question "Do you trust the Government?"
to which the answer is going to be "No".
Dr Forbes: With another system though;
compared with what other systems, this is the one we want to trust.
Professor Sasse: Trust is not on or off,
there are degrees and in a democracy it is absolutely right that
you should not trust Government blindly. However, if the trust
base between citizens and Government in general is very low, it
influences people's behaviour. If you look at the number of people
who are actually engaged in political processes, wanting to become
involved in Government, if you look at voting figures and so on,
those are also things that are connected to low trust in Government
and that is really not desirable.
Q458 Lord Norton of Louth: Moving
on to the RAE report's emphasis on the importance of public engagement
in policy formulation, it stresses that it is very important that
when discussing issues like privacy and related issues that there
should be some arrangements in place that actually facilitate
the involvement of citizens in policy formulation. What sort of
arrangements? How feasible is it to give them that? I can see
the principle but it is the operation of that principle. How does
one actually achieve that?
Professor Sasse: In my view Government
has recently been very fond of just holding consultations which
are effectively rubber-stamping, opinion-poll-type things. I do
not have a great deal of faith in those. If you contrast them
then with more detailed investigations where people actually have
a chance to discuss scenarios that personally concern them and
then to relate their decisions, what is reported is quite different.
It needs to be a more in-depth engagement. Say, for instance,
you were deciding to change the voting procedures, you cannot
just ask whether you would trust the Government to put in a proper
internet voting system. You would need to show people what it
actually means, what it would require them to do, what somebody
else can see, where it would take place. It needs to be a meaningful
consultation.
Q459 Lord Norton of Louth: I can
see your point and clearly it links back to your point about trust
because if people are involved in the processes, they discuss
privacy, it is their input and they feel they have had some say
in it, then presumably they are going to trust that mechanism
more. How does one have that wider consultation? If it is consultation,
it is normally the usual suspects who respond, it is not people
who are generally affected by these sorts of issues. How can we
actually engage people in the deliberation so they feel they are
involved and actually have a meaningful input?
Dr Forbes: This is a mass society problem
and the problems that the technology brings also bring you some
possibilities. I think the citizens are asked for data quite a
lot and you could arrange that at certain times when a citizen
is asked for data they are also asked other questions which would
explicitly be about the consultation, would raise the kind of
issues that people are concerned about and get their views on
them, which can be very simply organised. Consultation used to
be just insider groups frankly, but it is not necessarily now;
in the last ten years there have been a lot of opt-in possibilities
for consultation. However, there does need to be outreach work
to give people the opportunity and to say you are going to collect
this data but you are only going to do it if it is okay with you
and these are the kinds of issues we are thinking about and these
are the kinds of options that there may be and what are we missing
and what do you care about? Some people do care a lot about their
data and they will give you feedback and others will not and that
is also data. Low voting figures can also mean that actually people
think it is alright.
Professor Thomas: There is a real problem
in helping people to understand the potential for a current act
to cause future damage. I do not imagine that when the Netherlands,
back before the Second World War, decided to include religious
persuasion in their census data, they actually imagined they were
going to be invaded by the Nazis and that it would be used to
round up the Jewish population. Somehow you need to help people
to look ahead. It is not out of the question that a future Government
would decide that it was going to introduce a taxation regime
that discriminated against people who had not looked after their
health in the past, for example, and it is not out of the question
that they would use information from Tesco store cards in order
to gather that sort of data. I do not imagine that most people
signing up for a Tesco store card have in mind that that is a
risk that they are exposing themselves to. They might very well
decide to sign up anyway, but it does seem right that something
should be done at least to raise the level of awareness about
the potential when you can store such huge amounts of data and
search it so very, very easily.
| 
