Lawmakers weighing how to limit the spread of disinformation and control the power of Facebook and other social media giants are turning their attention to regulating the powerful artificial intelligence programs that drive them.
Washington has made little progress in stepping up regulation of big tech and has been hamstrung by First Amendment protections on free speech.
But Facebook whistleblower Frances Haugen told senators this month that the best way to do it was to focus on the powerful machine-learning architecture that promotes certain content.
The implications were spelled out on Tuesday, when leaked documents revealed how Facebook spent three years making misinformation and clickbait more prominent in users’ news feeds to keep them more engaged with its network.
It offers an alternative strategy for lawmakers wary of banning types of speech entirely.
Targeting algorithms – automated actions that rank content or direct users to relevant adverts, for example – could ensure that harmful content is spread less widely.
‘The algorithms driving powerful social media platforms are black boxes, making it difficult for the public and policy makers to conduct oversight and ensure companies’ compliance, even with their own policies,’ Senator Ed Markey, a Massachusetts Democrat, told Bloomberg.
Sen Ed Markey is one of the lawmakers who has set his sights on the algorithms that power social media giants in an effort to provide better oversight of such far-reaching technology
Facebook CEO Mark Zuckerberg consults with his legal team as he testifies at a House committee hearing in Washington as social media companies come under scrutiny
A Facebook whistleblower has claimed the company drives engagement by picking out content that is divisive in order to increase reactions from users
He is one of a number of lawmakers to have recently introduced bills that would ensure algorithms would operate in a more transparent manner.
That took on growing urgency with evidence provided by Haugen to a Senate subcommittee at the start of the month.
‘One of the consequences of how Facebook is picking out that content today is that it’s optimizing for content that gets engagement, or reaction,’ she told 60 Minutes ahead of her appearance.
‘But its own research is showing that content that is hateful, that is divisive, that is polarizing — it’s easier to inspire people to anger than it is to other emotions.’
Lawmakers have long discussed whether the huge power of companies such as Google and Facebook mean they should be broken up.
But Haugen’s testimony, outlining how unseen A.I. programs sent users radicalized information ahead of the Jan. 6 attack on the U.S. Capitol, has changed the direction of debate.
‘It’s heartening to see Congress finally beginning to focus on the heart of the problem,’ Rep. Tom Malinowski (D-N.J.), told the Washington Post recently.
‘The heart of the problem is not that there’s bad stuff posted on the Internet. It’s that social networks are designed to make the bad stuff spread.’
Even Facebook is coming around.
‘We need greater transparency,’ said Nick Clegg, vice president for global affairs at Facebook, on CNN’s ‘State of the Union recently.
‘[Our algorithms] should be held to account, if necessary by regulation, so that people can match what our systems say they’re supposed to do and what actually happens.’
Former Facebook employee-turned-whistleblower France Haugen gave evidence to British MPs this week after copying thousands of pages of internal research documents
The five Facebook emojis of ‘love,’ ‘haha,’ ‘wow,’ ‘sad’ and ‘angry’ were launched five years ago to give users an alternative way to react to content aside from the normal ‘like’
The firm’s algorithm, which decides what people see on a newsfeed, was allegedly programmed to use the reaction emoji as a sign to push more provocative content.
The five emojis of ‘love,’ ‘haha,’ ‘wow,’ ‘sad’ and ‘angry’ were launched five years ago to give users an alternative way to react to content aside from the traditional ‘like’.
But a ranking algorithm meant emoji reactions were treated as five times more valuable than ‘likes’, according to internal papers revealed by the Washington Post.
This idea behind this was that high numbers of reaction emojis on posts were keeping users more engaged – a crucial element to Facebook’s business model.
However, the company’s own researchers and scientists found that posts prompting angry reactions were far more likely to include misinformation and low-quality news.
One staffer allegedly wrote that favouring ‘controversial’ posts such as those making people angry could open ‘the door to more spam/abuse/clickbait inadvertently’.
Another is said to have replied: ‘It’s possible’. In 2019, its data scientists confirmed the link between posts sparking the angry emoji and toxicity on its platform.
This means Facebook stands accused of promoting the worst parts of its site for three years – making it more prominent and seeing it reach a much bigger audience.
It would have also had a negative effect on the work of its content moderators who were trying to reduce the amount of toxic and harmful posts being seen by users.
The discussions between staff were revealed in papers given to the Securities and Exchange Commission and provided to Congress by the lawyers of Haugen.
The whistleblower was in this week and said that Facebook was ‘unquestionably’ making online hate worse because it is programmed to prioritise extreme content.
She told MPs and peers that bosses at the firm were guilty of ‘negligence’ in not accepting how the workings of their algorithm were damaging society.
The American data scientist claimed the tech giant was ‘subsidizing hate’ because its business model made it cheaper to run angry and divisive adverts.
She said there was ‘no doubt’ the platform’s systems would drive more violent events because its most extreme content is targeted at the most impressionable people.
Haugen also issued a stark warning to parents that Instagram, owned by Facebook, may never be safe for children as its own research found it turned them into addicts.
She also told the joint committee on the draft Online Safety Bill that it was a ‘critical moment for the UK to stand up’ and improve social media.
The Bill will impose a duty of care on social media companies to protect users from harmful content and give watchdog Ofcom the power to fine them up to 10 per cent of their global turnover.
Facebook is currently battling a crisis after Haugen, a former product manager at the firm, leaked thousands of internal documents that revealed its inner workings.
Its founder Mark Zuckerberg has previously rejected her claims, saying her attacks on the company were ‘misrepresenting’ the work it does.