The Economist May 28th 2022 Britain 23InternetregulationHeadline removed for your own safety
B
ritain’sgovernmentlikestotrumpet
the benefits of free speech. Announcing
a bill designed to prevent stroppy students
“noplatforming” speakers at universities,
Boris Johnson told his millions of Twitter
followers last year: “Freedom of speech is
at the very core of our democracy.” The
Higher Education (Freedom of Speech) Bill
was included in the Queen’s Speech, which
sets out the government’s legislative agen
da, earlier this month.
Outside the lecture hall, though, Mr
Johnson’s government is accused of cen
sorship in its own right. Also in the
Queen’s Speech was the Online Safety Bill
(osb), a bumper piece of legislation that
will impose sweeping new obligations on
search engines, socialmedia sites, fo
rums, video sites and the like. Ministers
say it is “worldleading”. At 225 pages, with
194 separate clauses, and with up to 25,000
firms potentially affected, everyone agrees
it is ambitious. The arguments are about
its consequences.
To its defenders, the bill—whose ori
gins date back to Theresa May’s time as
prime minister—is designed to make Brit
ain “the safest place in the world in which
to use the internet”. The osbwill impose a
“duty of care” on tech firms, a concept that
began in healthandsafety legislation for
the workplace. Tech firms will be required
by law to protect their British users from
racism, death threats, sexual exploitation,
dodgy adverts and much, much more.
Any site likely to be accessed by chil
dren—which, in practice, means most of
them—coulddemandidentificationtoen
sure visitors are over 18. The biggest sites
will be required to scour everything their
users upload for potential illegality—and
perhaps also for things that, despite being
legal, are deemed “harmful” in Whitehall.
Penalties run to 10% of a firm’s worldwide
revenues, or even an outright ban.
Civil libertarians are not happy. A legal
opinion commissioned by Index on Cen
sorship, a charity, said the bill was likely to
fall foul of the freespeech provisions of
the European Convention on Human
Rights (to which Britain remains a signato
ry). David Davis, a Conservative mp, has de
scribed it as a “censor’s charter” that would
“strangle free speech online”.
The bill establishes several tiers of nas
tiness. The worst tier is reserved for things
that are against existing laws, such as as
sisting suicide, making threats to kill or as
sisting illegal immigration. The biggest
firms (the definition of “big” remains un
clear) will have to proactively purge such
things from their sites. The sheer size of
big platforms (500 hours of video are
uploaded to YouTube every minute) means
that it is not feasible for humans to check
every post. Firms will have to rely on auto
mated enforcement.
But algorithms are blunt tools, says
Mark Johnson of Big Brother Watch, which
campaigns on civil liberties. They often
struggle with nuance and context. “Is an al
gorithm going to be able to reliably tell the
difference between someone encouraging
suicide and someone with postnatal depression posting about feeling suicidal on
Mumsnet [a big web forum]?” With billions
of dollars potentially at stake, he says, the
risk is that firms will err heavily on the side
of caution, leading to overzealous blocking
of innocuous posts.
A second tier of content concerns posts
that are not in themselves illegal, but
which are deemed to be “harmful”. Exactly
what belongs in that category is unclear (it
is left for ministers to decide later). But the
government has talked about everything
from vaccine scepticism and bullying to
glorifying anorexia and hurling racist in
sults at England’s football team. Sites will
be required by law to minimise the chance
that children see such posts. For everyone
else, they will have to make an active deci
sion about whether to block or downplay
such content, or to carry on promoting and
recommending it as they would anything
else on their products.
Campaigners argue that these “legal but
harmful” provisions amount to censorship
by the back door, and create an entirely
new category of speech in law. Ruth
Smeeth, a former mpand the boss of Index
on Censorship, calls them a “clusterfuck”.
The government has tried to reassure
doubters by pointing out that tech firms
will be free to leave such posts up if they so
choose. Ms Smeeth is dismissive: “Can you
imagine the political pressure on any plat
form that says publicly they’re okwith this
stuff?” Those that decide to suppress it will
again depend on idiotsavant algorithms.
There are other doubts. A string of new
communications offences rely on subjec
tive definitions of psychological harm.
Jonathan Hall, the government’s indepen
dent reviewer of terrorism legislation, is
unhappy with similar vagueness around
definitions of terrorism. Some police offi
cers worry that firms could destroy digital
evidence of crime by deleting posts. But
support in Parliament remains strong. Last
year Sir Keir Starmer, the Labour leader,
castigated the government for not moving
more quickly. The bill looks likely to make
it onto the statute books. Online, off limits
That could have international repercus
sions. Tech firms may choose to try to ap
ply the new laws in Britain only. If they de
cide that it is too difficult to create a new
set of rules for a subset of their users, one
option is to apply at least some of the osb’s
provisions to their services anywhere.
There is precedent: last year Britain
brought in the Age Appropriate Design
Code, which prescribes stricter privacy for
children online. Google, TikTok and others
made worldwide changes as a result. An
other possibility is that some foreign
firms, particularly smaller ones,maystop
serving Britain entirely. In variousways,
the price of safety may be silence.nThe Online Safety Bill promises huge changes to what Britons can do online