Facebook’s AI is removing just TWO PER CENT of hate speech posts
Facebook’s AI is removing just TWO PER CENT of hate speech posts that break social media giant’s own rules, with insiders saying the software is mistaking cock fights for car crashes, and failing to spot shooting clips
Facebook executives have long boasted about the success of their AI systemsMark Zuckerberg told Congress last year they had ‘built really sophisticated systems’ to combat hate speechYet internal documents obtained by The Wall Street Journal showed significant concerns about the ability to detect problematic contentCockfights were mistakenly labelled as car crashes, and videos of murders were remaining on the site for several hoursSinger Selena Gomez became so concerned at the level of abuse that she wrote to Zuckerberg and Sheryl Sandberg, the chief operating officerGomez, in a series of emails to them, said she was concerned people could be seriously harmed or even killedThe internal documents were obtained from Frances Haugen, the project manager who testified before Congress on October 5
<!–
<!–
<!–<!–
<!–
(function (src, d, tag){
var s = d.createElement(tag), prev = d.getElementsByTagName(tag)[0];
s.src = src;
prev.parentNode.insertBefore(s, prev);
}(“https://www.dailymail.co.uk/static/gunther/1.17.0/async_bundle–.js”, document, “script”));
<!–
DM.loadCSS(“https://www.dailymail.co.uk/static/gunther/gunther-2159/video_bundle–.css”);
<!–
Facebook‘s Artificial Intelligence programs detect and remove as little as two per cent of hate speech posted on the platform – despite promises from Mark Zuckerberg that it was the future for content moderation.
Internal documents obtained by The Wall Street Journal showed the scale of the problem with the social media giant’s machine-learning software, while senior figures at the tech giant were insisting publicly that their AI schemes were efficient and effective.
In July 2020, Zuckerberg told Congress: ‘In terms of fighting hate, we’ve built really sophisticated systems.’
Two years before he had told a Senate committee that he was optimistic that within five to 10 years, Facebook would have the AI tools to detect most hate speech.
‘Over the long term, building AI tools is going to be the scalable way to identify and root out most of this harmful content,’ he said.
Mark Zuckerberg, pictured with Sheryl Sandberg, the chief operating officer, told Congress the AI systems Facebook developed were ‘really sophisticated.’ Yet internal documents showed that the systems only detected an estimated two per cent of all hate speech
Yet in mid 2019, a senior engineer and research scientist warned that AI was unlikely to ever be effective in internal documents now uncovered by the WSJ.
‘The problem is that we do not and possibly never will have a model that captures even a majority of integrity harms, particularly in sensitive areas,’ he wrote.
He estimated the company’s automated systems removed posts that generated two per cent of the views of hate speech on the platform that violated its rules.
‘Recent estimates suggest that unless there is a major change in strategy, it will be very difficult to improve this beyond 10-20% in the short-medium term,’ he wrote.
In 2018 engineers became concerned that videos of cockfighting were being noted by the system as car crashes.
They tried to tweak the system to allow scenes that did not show severely injured birds, but the AI proved incapable of detecting the variation, despite being fed clips of varying degrees of animal abuse to try and teach it to identify what broke the rules.
The documents also detailed how, in March 2019, the AI system failed to detect a live-stream of a mosque shooting in Christchurch, New Zealand, which killed 51 people.
The footage remained online for hours after the attack. This was because of a glitch that mean Facebook’s AI struggled to register first-person shooter videos – those shot by the person behind the gun.
Cockfights were mistakenly flagged as car crashes, and the AI system was unable to differentiate between severely injured animals and less hurt ones
Selena Gomez in 2016 visited Facebook’s headquarters. But several years later she complained to Zuckerberg and Sandberg about the threat posed by hate speech online
Singer Selena Gomez – who in 2016 visited Facebook’s headquarters to celebrate her being the most followed person on Instagram, which Facebook owns – wrote to Sheryl Sandberg, Facebook’s chief operating officer expressing her concern.
She pointed out that, beneath one of her posts, someone had written: ‘Go kill yourself.’
She told Zuckerberg and Sandberg that the company had a ‘serious problem’ with hate, misinformation, racism and bigotry.
Sandberg responded that Facebook’s AI had detected 91 per cent of the 1.5 million posts it removed for violating its rules against using symbols or phrases from hate groups.
Gomez replied that Sandberg had not responded to her broader questions, and sent screenshots of Facebook groups that promoted violent ideologies.
‘You refuse to even mention, let alone address, the problem Facebook has with white supremacists and bigots,’ Gomez wrote in an Oct. 10, 2020, email to Ms. Sandberg and other executives, adding that there were plenty of Facebook groups ‘full of hate and lies that might lead to people being hurt or, even worse, killed.’
Andy Stone, a Facebook spokesman, said the information from the 2019 presentation was outdated.
Zuckerberg has insisted publicly that AI is solving many of Facebook’s hate speech problems
But in March, another team of Facebook employees reported that the AI systems were removing only 3-5 per cent of the views of hate speech on the platform, and 0.6% of all content that violated Facebook’s policies against violence and incitement.
The internal memos came as Facebook was publicly insisting that AI was working well.
The Silicon Valley firm states that nearly 98 per cent of hate speech was removed before it could be flagged by users as offensive.
Yet critics say that Facebook is not open about how it reached the figure.
‘They won’t ever show their work,’ said Rashad Robinson, president of the civil rights group Color of Change, which helped organize an advertiser boycott of Facebook last year due to what it called the company’s failure to control hate speech.
He told the paper: ‘We ask, what’s the numerator? What’s the denominator? How did you get that number?
‘And then it’s like crickets.’
Facebook says five out of every 10,000 content views contained hate speech, an improvement from roughly 10 of every 10,000 views in mid-2020.
‘I want to clear up a misconception about hate speech on Facebook. When combating hate speech on Facebook, bringing down its prevalence is the goal,’ tweeted Guy Rosen, the vice president of integrity, on October 3.
‘The prevalence of hate speech on Facebook is now 0.05%, and is down by about half over the last three quarters.
‘We can attribute a vast majority of the drop in prevalence in the past three quarters to our efforts.’
Frances Haugen, a former product manager hired by Facebook to help protect against election interference, leaked the documents to The Wall Street Journal. She testified before Congress (pictured) on October 5
Facebook says it has spent about $13 billion on ‘safety and security’ since 2016, or nearly four per cent of its revenue in that time.
Review of hate speech by human staff was costing $2 million a week, or $104 million a year, according to an internal document covering planning for the first half of 2016.
‘Within our total budget, hate speech is clearly the most expensive problem,’ a manager wrote.
The documents published by The Wall Street Journal were leaked by Frances Haugen, 37, who left Facebook in May after nearly two years.
Haugen, a former product manager hired by Facebook to help protect against election interference, testified before Congress on October 5.
She argued for greater governmental oversight of tech companies, insisting that executives knew the harm done but did little to stop it.
