Google is changing the way its core search engine works to help stop the spread of fake news and hate speech.
The changes involve different measures for ranking sites and people checking results are accurate.
In a blog, Google said the changes should thwart attempts to abuse its algorithms that let extremists promote their content.
Google was criticised last year for giving prominence to groups seeking to deny that the Holocaust took place.
Rate and replace
Ben Gomes, a vice-president of engineering at Google’s search division, said it was making “structural” changes to tackle the new ways people had found to trick its algorithms.
In particular, he said, many groups and organisations were using “fake news” to help spread “blatantly misleading, low quality, offensive or downright false information”.
To combat this, he said, Google had added new metrics to its ranking systems that should help to stop false information entering the top results for particular search terms.
In addition, he said, it had updated the guidelines given to the thousands of human raters it used to give feedback on whether results were accurate.
The guidelines included examples of low quality and fake news websites, said Mr Gomes, to help them pick out “misleading information, unexpected offensive results, hoaxes and unsupported conspiracy theories”.
Analysis: Rory Cellan-Jones, BBC technology correspondent
Google has done its best to play down the extent of fake news and hateful material – or what it prefers to call “low quality content” – in search results.
The company keeps repeating that this affects only 0.25% of queries.
But the fact that searches such as “Is Obama planning a coup?” – or even “Who invented stairs?” – produced such questionable results meant it had to act.
These searches threw up a prominent “snippets” box telling you that, yes, President Obama was planning a coup, or that stairs had been invented in 1948.
Now both boxes have gone, and Google’s almighty algorithm has been tweaked so that such content is less likely to rise to the top.
What’s interesting is that a company that has put such faith in technology solutions is turning to 10,000 humans to try to make search a better experience.
This giant focus group, which tests out changes in the search algorithm, has been told to pay more attention to the source of any pages rated highly in results, looking round the web to see whether they seem authoritative and trustworthy.
Questions are bound to be raised about whether this panel, which Google says is representative of its users, is impartial and objective.
Google’s Ben Gomes, a veteran who’s been wrestling with the intricacies of search since arriving as one of the earliest employees, believes it is now on the path to getting this right.
But with so many people trying to game the system, the battle to make search true and fair will never be over.
Google also planned to change its “autocomplete” tool, which suggests search terms, to allow users to more easily to flag up troubling content, he said.
Danny Sullivan, founder of the Search Engine Land news site, said the changes made sense and should not be taken to suggest that Google’s algorithms were failing to correctly index what they found online.
“It’s sort of like saying that a restaurant is a failure if it asks for people to rate the food it makes,” he said.
“The raters don’t rank results,” said Mr Sullivan.
“They simply give feedback about whether the results are good.
“That feedback is then used to reshape the algorithms – the recipes, if you will -that Google uses.”
Wikipedia’s co-founder Jimmy Wales is planning a news service that combines the work of professional journalists and volunteers.
His goal is for Wikitribune to offer “factual and neutral” articles that help combat the problem of “fake news”.
The service is intended to be both ad-free and free-to-read, so will rely on supporters making regular donations.
One expert said it had the potential to become a trusted site, but suggested its influence might be limited.
Wikitribune shares many of the features already found in Mr Wales’s online encyclopaedia, including the need for writers to detail the source of each fact and a reliance on the public to edit articles to keep them accurate.
However, while anybody can make changes to a page, they will only go live if a staff member or trusted community volunteer approves them.
The other big difference is that the core team of writers will be paid, although there may also be instances in which a volunteer writes the initial draft and then a staff member edits it.
A demo version of the site, seen by the BBC, declared “the news is broken and we can fix it”.
Mr Wales explained that he believed the advertising-based model used by most of the media had led it to “chase clicks”, which affected standards.
“I think we’re in a world right now where people are very concerned about making sure we have high quality fact-based information, so I think there will be demand for this,” he told the BBC.
“We’re getting people to sign up as monthly supporters and the more monthly supporters we have the more journalists we can hire.
“In terms of minimums, if we could only hire two journalists then it would be a blog and not really worth doing.
“But I would love to start with a lot more – 10 to 20.”
The director of Harvard University’s Nieman Journalism Lab, however, suggested Wikitribune’s crowdfunded model might limit its potential.
“There are a variety of people who – if it does this right – will view it as a trusted platform,” commented Joshua Benton.
“But another 10 to 20 people are not going to ‘fix the news’.
“There’s certainly a model for non-profit news that can be successful if it’s done on a relatively small scale and produces a product that is unique enough.
“But I have a hard time seeing this scale up into becoming a massive news organisation.”
Mr Wales said he would be “100% hands-on” with the project in its early stages and would be likely to serve as Wikitribune’s chief executive for at least a year.
Other advisors to the scheme include:
Silicon Valley venture capitalist Guy Kawasaki
Journalism lecturer Prof Jeff Jarvis
US law professor Larry Lessig
Model/actress Lily Cole
Although staff members will decide the topics that get written about on a day-to-day basis, funders will get to influence the contents.
“If you can get together a certain number of people who are interested in Bitcoin [for example] and you flag that when you sign up as a monthly supporter, then we’ll hire a Bitcoin person to do the beat full-time,” Mr Wales explained.
“So, it’s the monthly supporters who will be able to determine what are the topics we are going to cover.
“But it is going to be neutral. They can’t pick their favourite hack, who pumps forward their agenda.
Jimmy Wales’s interest in news media is nothing new. For years he has expressed concern about how to guarantee the future of quality journalism, and even been talked of as a potential investor in existing media companies.
But when I spoke to him yesterday, it was clear that there was something new – or rather three things – that finally turned his long-standing interest into the reality of Wikitribune.
The first is what we call fake news. Fake news is a multi-faceted thing, and not altogether new; but it is undoubtedly the case that the deliberate, viral spreading of misinformation, either for commercial or political ends, has radically spiked around some of the big news events of the past year. Moreover, efforts to tackle it have often been pathetic thus far, and less often successful. This really irks Wales, and quite right too.
The second recent development is the radical shift in online advertising, where the strength of Facebook and Google – who are gobbling up ever more digital advertising dollars – is creating a race to the bottom. “I’m very concerned by the advertising-funded model, which is creating a lot of clickbait”, he told me.
And third, mounting evidence that people are willing to pay for high-quality news. Wales cited New York Times subscriptions and Guardian membership. He might also have mentioned the Financial Times.
To address the first two of these developments, Wales is looking to the third: he wants to get users to pay for news, and then play a hugely active role in determining its focus.
I argued recently that charity is a poor basis for journalism; much better to get users to pay. Wales agrees. He thinks by asking users to invest financially, they are more likely to invest emotionally too. I think he’s onto something – but it depends on whether what they get as a result is worth investing in.
And here’s the rub.
Nobody, but nobody, has as much credibility as Jimmy Wales when it comes to proving the wisdom of crowds exists online, or that the sheer scale of the open web allows knowledge to be shared and chronicled. But can the spirit of public participation that drove an online encyclopaedia also drive online news?
We don’t know, because the fascinating thing about Wikitribune – whose name is redolent of old newspapers – is that it isn’t just reinventing the commercial model for journalism: it’s reinventing the editorial one too.
The function of an editor is mainly to select what to put in and what to leave out. News has traditionally been selected by editors, who are gatekeepers and curators. But Wales, who is the founding editor of this publication, doesn’t see it like that. “It’s more a management role than editorial vision or pursuing an agenda,” he told me.
This is fascinating, and sounds very similar indeed to the arguments Ben Smith, the editor-in-chief of Buzzfeed, deployed when he published a dossier on Donald Trump.
It is a curious fact that while setting out to save journalism, Jimmy Wales is abolishing one of its most traditional roles. He would argue, of course, that he is giving power back to the audience in a way they have never had before: letting them be the editors, rather than pompous blowhards who think they know best.
This might discomfort many a grandee in the news profession; but if the man from Wikipedia provides a business model that sustains top notch reporting, they might thank him eventually.
The Facebook Live post was widely reported by Thai media, and went viral on social media, BBC Thai editor Nopporn Wong-Anan reports.
In a statement, a Facebook spokesperson said: “This is an appalling incident and our hearts go out to the family of the victim. There is absolutely no place for content of this kind on Facebook and it has now been removed.”
YouTube said it had taken down the video within 15 minutes of being told of its presence by the BBC.
Its statement read: “YouTube has clear policies that outline what’s acceptable to post and we quickly remove videos that break our rules when they’re flagged.”
Shortly before the BBC alerted YouTube, the video was showing 2,351 views.
Thai social media users reacted with anger to the footage, while offering condolences to the family of the girl, our correspondent says.
Devastated relatives of the child, including the mother, picked up the body of the girl and her father from hospital on Tuesday.
Following the US killing, Facebook said it was “constantly exploring ways that new technologies can help us make sure Facebook is a safe environment”.
This latest atrocity comes less than a fortnight after a US man bragged on Facebook Live about his murder of a 74-year-old man in Cleveland, having also posted a video of the killing to the social network.
The platform’s chief, Mark Zuckerberg, subsequently acknowledged he had “a lot of work” to do after it emerged the murder clip had remained online for more than two hours despite Facebook having received complaints in the meantime.
Engineer Harry Huskey, who helped build many of the first ever computers, has died aged 101.
Dr Huskey was a key member of the team that built the Electronic Numerical Integrator and Computer (Eniac) which first ran in February 1946.
Eniac is widely considered to be one of the first electronic, general purpose, programmable computers.
Dr Huskey also helped complete work on the Ace – the Automatic Computing Engine – designed by Alan Turing.
The Eniac was built at the University of Pennsylvania in the 1940s and, once complete, was more than 100ft (30m) long, weighed 30 tonnes, used 18,000 valves and 1,500 relays. Programming the massive machine to do different computational tasks involved rewiring its various units. Eniac was built to calculate the trajectory of shells for the US army.
Dr Huskey became involved with the development effort to create Eniac soon after joining Pennsylvania to teach mathematics to Naval recruits. His task was to make the punched card reader for the machine work and to write technical manuals describing how to operate it.
After the war, Dr Huskey travelled to the UK to help Alan Turing refine and complete the Ace. This was built at the National Physical Laboratory and in 1950, when it ran its first program, it was the fastest computer in the world.
He also helped design and build two other machines – the Swac (Standards Western Automatic Computer) and the G-15 which, despite weighing almost a tonne. was known as a personal computer because it could be operated by one person.
Dr Huskey spent his entire academic career involved with computing teaching at the University of California, Berkeley and was one of the founders of the computer science faculty at UC Santa Cruz.
“Harry basically lived through and participated in the entire span of the history of electronic computing,” Dag Spicer, a curator at the Computer History Museum, told the New York Times.
The Brain Tumour Charity has said there is insufficient scientific evidence linking mobile phone use with brain tumours, following a court ruling.
The Italian court, in Ivrea, agreed that a man’s brain tumour was linked to his mobile phone use.
It awarded Robert Romero 500 euros (£418/$535) a month in compensation.
He had claimed that using his business mobile phone for three or four hours a day, over a period of 15 years, led to the growth of the benign tumour.
The money will be paid by a body established to compensate people for work-based injuries.
There could yet be an appeal against the ruling, and the legal reasoning behind the judge’s decision is not due to be released for at least a few days.
“We know that many people are concerned about a possible connection between mobile phone use and the development of brain tumours,” said Dr David Jenkinson, chief scientific officer for the Brain Tumour Charity.
“However, the global research projects that have been conducted so far, involving hundreds of thousands of people, have found insufficient evidence that using a mobile phone increases the risk of developing a brain tumour.”
The decision of the court did not change the evidence, he added.
“Of course, it is right that researchers continue to explore whether any such link exists,” said Dr Jenkinson.
Mr Romero, whose profession was not reported, said he wanted people to be more aware about mobile phone use but did not want to “demonise” the devices.
His lawyer, Stefano Bertone from the law firm Ambrosio and Commodo, told the BBC his client currently has no plans to sue any of the handset manufacturers or the mobile phone industry itself.
He added that the firm has other cases in other parts of Italy.
“We have also been approached by an interesting number of people in the last 24 hours saying they have experienced the same kind of thing. And they can show they have accumulative use of mobile phones that’s exceeding 1,000 hours,” he said.
“No-one can pretend with definitive certainty to assess a legal case. Most opponents say there is no scientific certainty so therefore it is not true. That is not the case.”
Preliminary findings released in 2016 suggested a “low incidence” of brain and heart tumours in male rats exposed to doses of radiofrequency radiation totalling up to nine hours a day over a two-year period.
However, as it is not finished, the study has not yet been scrutinised by other scientists, a process known as peer reviewing, which is generally considered an essential stage of evaluating research.
A payment card featuring a fingerprint sensor has been unveiled by credit card provider Mastercard.
The rollout follows two successful trials in South Africa.
The technology works in the same way as it does with mobile phone payments: users must have their finger over the sensor when making a purchase.
Security experts have said that while using fingerprints is not foolproof, it is a “sensible” use of biometric technology.
Mastercard’s chief of safety and security, Ajay Bhalla, said that the fingerprint technology would help “to deliver additional convenience and security. It is not something that can be taken or replicated.”