The EconomistAugust 10th 2019 Britain 211F
or a time, the humans seemed to be
winning. The Metropolitan Police’s
squad of “super-recognisers” was lauded
for its uncanny ability to recall faces in vid-
eo footage. The officers spotted sex-offend-
ers in crowds of thousands and nabbed a
thief who had pinched more than £100,000
($122,000) of luxury goods. Technology
was not nearly as reliable. One of the super-
recognisers identified 180 of the 4,000 sus-
pects captured on camera during riots in
2011, whereas software spotted only one.
“Computers are no match for the super-re-
cognisers,” boasted the unit’s boss.
Now the computers are fighting back.
Many of the 43 police forces in England and
Wales are experimenting with algorithmic
technology that could render the copper’s
nose redundant. Several use programs to
predict where and when crimes are likely
to occur. Cambridge University helped
Durham Constabulary design an algorithm
to estimate the likelihood of a suspect reof-
fending. It helps the authorities decide
whether someone should be granted bail or
qualify for rehabilitation as an alternative
to prosecution. At least one force is keen to
install microphones on “smart lamp-
posts” to gather intelligence in crowds.
Even the cherished super-recognisers will
be outdone once facial-recognition algo-
rithms improve, predicts Rick Muir of the
Police Foundation, a think-tank.
Not everyone is pleased. On July 3rd aca-
demics published a critical assessment of
Scotland Yard’s pilots of automatic facial-
recognition technology, querying their le-
gal basis and casting doubt on whether
people caught on camera could be said to
have given their informed consent. Judges
in Cardiff are weighing the lawfulness of
similar trials by South Wales Police. And in
June the Law Society, which represents so-
licitors, raised concerns about the “general
and concerning lack of openness or tran-
sparency” in the police’s use of algorithms.
Several wonk shops are being set up to
examine the ethics of algorithmic technol-
ogy, including one at Oxford backed by a
£150m donation from Stephen Schwarz-
man, the boss of Blackstone, a private-equ-
ity firm. The Centre for Data Ethics and In-
novation, a new government-funded
agency, is likely to propose a code of con-
duct to regulate cops’ use of technology,
says Roger Taylor, its chairman, who ac-
knowledges the need to act “very quickly”
to close any gaps in oversight.Critics make four arguments. First, the
technology does not work terribly well. In
the pilots in London, only eight of the 42
matches made by facial-recognition soft-
ware were correct. Second, the systems are
a disproportionate response to crime. In an
era when data-protection regulations gov-
ern the mailing list of a pizza joint, civil-
liberties campaigners question why the
national police database holds 12.5m im-
ages in its gallery—including images of an
undisclosed number of people who have
neither been charged with an offence nor
consented to the use of their pictures.
Third, it could prove discriminatory.
Since some facial-recognition technology
is best at identifying white faces, it could
throw up more erroneous “matches” for
non-white people, making them more like-
ly to be the subject of unwarranted police
attention. Finally, it risks compromising
the principle that justice must be seen to be
done. If suspects cannot understand how
an algorithm reached a decision, they
might find it harder to challenge.
Yet none of these hurdles is insur-
mountable. The technology will improve.
Britons already accept lots of surveillance:
although most people do not shoplift, they
are used to being monitored by cctvcam-
eras. A poll published in May suggests most
Londoners are happy for the police to use
facial-recognition software, especially to
spot serious criminals. A powerful regula-
tor ought to be able to strike the right bal-
ance and allay fears of bias.
And although humans can give reasons
for their decisions, there is plenty of evi-
dence suggesting they are influenced by
unconscious biases, points out Lawrence
Sherman of Cambridge University. It ought
to be easier to scrutinise and challenge the
processes of one algorithm than the deci-
sions of thousands of cops and judges.
“There’s nothing less transparent than the
human mind,” says Mr Sherman. 7Justice by algorithm is in the dockPolicing and technologyHold on a
millisecond
I
n his firstspeech as the leader of the
Conservative Party, Boris Johnson ex-
plained his governing philosophy: “Dude,
we’re going to energise the country!” His
first two weeks in office have seen him do
his best to live up to the laugh line. He has
criss-crossed the country, leaving a trail of
spending commitments in his wake. Live
in a left-behind town? He will spruce up
your hospital. Based in a growing Northern
metropolis? Faster, more frequent trains
for you. Concerned by rising violent crime?
Worry not: 20,000 extra police officers are
on their way.
Although the government insists there
are no plans to call an election, it appears
on a collision course with Parliament over
Brexit and has a dwindling parliamentary
majority—down to just one after defeat by
the Liberal Democrats at a by-election in
Brecon on August 1st—meaning many ex-
pect a vote within the year. The buffet of
policies laid out by Mr Johnson suggests he
wants to rebuild the electoral coalition the
Leave campaign brought together, includ-
ing not just Tory Brexiteers, but also left-
leaning, culturally conservative types in
the Midlands and the north (see Bagehot).
Downing Street has attempted to claim
that an economic ideology lies behind
these announcements, labelling Mr John-
son’s promise of both tax cuts and higher
spending as “boosterism” (rather than the
more obvious “cakeism”). He has promised
£3.6bn ($4.4bn) for 100 poor towns, £2.1bn
for the National Health Service and £1.1bn
for hiring new police officers. He has
voiced support for new railways to join up
the north which could cost a whopping
£39bn. The flurry of spending commit-
ments provides an indication of how the
new prime minister will govern and, per-
haps, campaign.
He has moved quickly to cauterise is-
sues that may harm the Conservatives on
the doorstep. The main teachers’ union
spent £325,000 at the last election, more
than both the ukIndependence Party and
the Greens, in a successful campaign for
more education funding. Mr Johnson has
promised to reverse recent school cuts by
2022-23. Health care is one of the few policy
areas where the public still trusts Labour
more than the Tories. Dominic Cummings,
Mr Johnson’s chief adviser, has com-
plained that few Conservative mps have
faced up to the importance of the nhsin
British politics. On August 5th the primeThe new prime minister promises
something for every voterPublic servicesBoris gets out his
wallet