(Reuters) - In the closing days of the British election campaign, Prime Minister Theresa May's bid to blame internet companies for fostering "safe places" where extremists breed has backfired.
The government's moves to make an example of Silicon Valley have met with widespread criticism from rival politicians and security experts for their lack of workable policy proposals.
Her comments in the wake of this weekend's London attacks drew attention to her own record of cutting police budgets and failing to crack down on home-grown jihadists in a campaign that polls nonetheless show she is likely to win.
There is mounting evidence that jihadists responsible for recent attacks in London and Manchester have relied largely on low-tech techniques rather than social media propaganda and encrypted communications tools that May has singled out.
The Conservative government has sought to enlist public opinion to force internet companies to work more closely with the government rather than proposing new legislation or policies to assert greater control over the web.
In recent months, the government has lambasted the big Silicon Valley companies - Google, Facebook and Twitter - for not quickly enough taking down extremist propaganda used to recruit "lone wolf" attackers.
"The companies should accept their responsibility in relation to what is being put across their platforms," May said in a campaign speech in central England on Tuesday.
"Because quite frankly we see hateful ideology being spewed across their platforms by the extremists that can lead to terrorism and we don't want to see a safe space online for the terrorists to be planning their attacks."
After resisting more forceful action for years, the internet firms say they are now making heavy investments and employ thousands of people to take down hate speech and violent content on their platforms, while acknowledging they continue to struggle to identify replacement accounts that quickly reappear.
There is evidence their efforts are working. Extremist monitoring group Site Intelligence Group said a pro-Islamic State tech group claimed to have generated 11,000 social media accounts for IS supporters in May, down from 15,000 in April, marking a continuing decline over the past year.
Google said it now manages to review 98 percent of complaints about videos on its YouTube service within 24 hours and has a "zero tolerance" policy toward extremist content.
"We want to make sure that terrorists do not have a voice and cannot spread extremist material on our services," Google spokesman Peter Barron said.
Twitter said that in the second half of 2016, it suspended 376,890 accounts for promoting what it labels as terrorism. Three-quarters of these violations were detected automatically, while most of the rest were from complaints by Twitter users.
In that time frame, Twitter said it received 681 requests from UK law enforcement for user data tied to 1,017 accounts.
"Terrorist content has no place on Twitter," Nick Pickles, the company's UK head of public policy, said in a statement.
Facebook declined to comment.
"Regarding the internet, much has changed since 2014. Big social media platforms have cracked down on jihadist accounts," Kings College extremism expert Peter Neumann said in response to May's Internet crackdown speech.
Widely used messaging services such as WhatsApp and Telegram secure private messages with end-to-end encryption, which means that messages can only be seen by the sender and the recipient.
Home Secretary Amber Rudd has called for a ban on encrypted messaging services, saying that there should be "no place for terrorists to hide".
Her assertions have been met with criticism for seeking to roll back privacy protections for consumers that are key for secure online banking or shopping, and because such moves have been shown over the years to encourage extremists to switch to new tools available in darker corners of the internet.
Technology experts complain that weakening encryption by creating backdoors for government will inevitably open up online messaging services to bad actors, leaving the phones, computers and other devices the world relies on far more vulnerable.
Security expert Neumann emphasised how few people are radicalised exclusively online and that the London attackers appeared to be part of a well-known jihadist network under hate preacher Anjem Choudary, who operated freely in Britain until 2015.
"Blaming social media platforms is politically convenient but intellectually lazy," Neumann said in a widely quoted tweet.
Reporting by Eric Auchard and Michael Holden; editing by Mark Heinrich