Striking the Right Balance between Privacy and Public Protection

[Ross Anderson]

Richard said:

> Worry, therefore, not so much whether someone can find your home address
> from a CD, but whether someone can pick your details from anywhere and use
> those to register a pre-pay mobile which they intend to use in connection
> with a "serious crime" - say drugs trafficking - which will subsequently
> be traced to that mobile phone and the mobile phone company will provide
> the police with the registered name and address ... YOUR name and address
> ... so causing you to become "reasonably suspected of being involved with"
> that crime and in consequence subject to all the other forms of intrusion
> that we have been discussing these recent weeks.

It can be much worse than just intrusion. You might easily end up
being convicted of a crime you did not commit. Most of the cases in
which I've acted as an expert for the defence, and in which I believed
that the accused really were innocent, hinged on computer evidence,
blindly trusted by the police, and padded out into a story with
assorted circumstantial tidbits. Historically, miscarriages are very
often associated with the one-sided use of expert evidence - whether
of a complainant's computer experts, or police fingerprint experts,
or Home Office forensic chemists, who had become complacent and
careless and whose evidence could not be challenged effectively.

The victims in such cases are typically the disadvantaged - old women,
ethnic minorities or even police officers (who are handicapped by the
relative vigour with which the PCA pursues complaints against junior
policemen). People on this list will mostly have heard of the Munden
case; there was also the McKie case, which demolished the fifty-year
old consensus that fingerprint evidence was completely reliable (see
first item below).

People who've not experienced the criminal justice system first-hand
would be surprised to know how little solid evidence is often needed
to send someone down.

The situation has recently become a lot worse since 

(a) the government removed the old PACE provision that computer
evidence had to be accompanied by a certificate to the effect that the
computer was working properly at the time. It was this requirement
that gave the defence the entry point to challenge prosecution cases
based on computer evidence that was erroneous or forged

(b) the Legal Aid Board is generally less willing to contemplate
paying defence solicitors to emply experts. Recently, for example,
expert help was refused in a case in which someone was prosecuted for
denying two Barclaycard transactions, on the grounds that the
Barclaycard now contains a chip and so it must be secure!

This change in the balance of arms between the prosecution and the
defence may be convenient for the Home Office in the short term. But
there is no free lunch here. I fear that, combined with greater access
to (and reliance on) a great variety of data from all sorts of systems
that were never designed to provide evidence, it is storing up serious
problems.

Simon, I am concerned that the wholesale use of data from ISPs and
telcos and the like will mean that prosecutors rely on much more
computer evidence, from a greater range of systems - and of dodgier
systems - than has been the case in the past, while defendants will be
less able to make credible challenges to this evidence (even when it's
wrong) and judges will be less prepared to listen to such challenges
as can be mounted.

If no compensating changes are made to the system, then sooner or
later there will be a series of high-profile miscarriages or other
abuses that will force a change. In the meantime, courts will slowly
lose confidence in expert evidence (see second and third items below),
while innocent people will get hurt with increasing frequency. Public
confidence in the system will be slowly undermined. Is this really in
the public interest?

Ross


***********


EVIDENCE
Criminal evidence; Fingerprints; Forensic evidence; Miscarriage of justice; 
Police; Scotland

"Fingerprint killer" takes on the police. By Dan McDougall and John 
Robertson. Scotsman, August 15, 2002, 1 Also reported in Herald, August 15, 
2002, 1

David Asbury convicted of murdering Marion Ross and sentenced to life in 
prison, has had his conviction squashed after the court accepted legal 
arguments proving the Scottish Criminal Records Office's fingerprint 
evidence against Mr Asbury was unreliable. Mr Asbury will now pursue a 
civil case against Strathclyde Police.

See also 

   http://news.bbc.co.uk/1/hi/uk/343940.stm

for the news story, and

   http://www.onin.com/fp/problemidents.html

for the full technical detail.

***********

CRIMINAL EVIDENCE
Admissibility; Ear; Expert evidence; Identification; Murder

R. v Dallagher (Mark Anthony) Times, August 21, 2002  2000/5024/Z2 Court of 
Appeal (Criminal Division). Judgment date: July 25, 2002

A conviction for murder, imposed after the defendant was identified on ear 
print evidence, was quashed and a retrial ordered, in the light of fresh 
evidence which questioned the safety of using ear print evidence alone in 
the identification of suspects.

Appeal against conviction allowed

***********

Jan 17th 2002
From The Economist print edition

A judge has ruled that fingerprint evidence is scientifically unreliable.

 "BOOK him," says the crusty sergeant at the local precinct, and another
villain is led off to be fingerprinted. For nearly a century, suspects have
lived in fear that matches to their prints will turn up at the scene of the
crime. Such fears have now been reduced a little. In a ruling on January
7th, Louis Pollak, a federal judge in Pennsylvania, decided that
fingerprint evidence was unreliable. He declared that he will no longer
allow fingerprint examiners testifying in his courtroom to tell juries
point blank whether prints from defendants and those collected at crime
scenes do or do not match. Instead, they will have to present evidence to
persuade a jury that they are the same or, as the case may be, are not.

This is the first ruling of its kind in the American courts, although
fingerprinting evidence has been open to such a challenge for years. In the
1990s America's Supreme Court deemed it the responsibility of federal
judges to insist that expert witnesses testify about the reliability of a
forensic-scientific method only if the method in question has been tested
so that the range of its error rate is known. Fingerprinting experts have
long claimed that their error rate in matching prints is but without
any supporting evidence.

...

***********