TED 2018: Fake Obama video creator defends invention

A researcher who created a fake video of President Obama has defended his invention at the latest TED talks.

The clip shows a computer-generated version of the former US leader mapped to fit an audio recording. Experts have warned the tech involved could spark a “political crisis”.

Dr Supasorn Suwajanakorn acknowledged that there was a “potential for misuse”.

But, at the Vancouver event, he added the tech could be a force for good.

The computer engineer is now employed by Google’s Brain division. He is also working on a tool to detect fake videos and photos on behalf of the AI Foundation.

Damage risk

Dr Suwajanakorn, along with colleagues Steven Seitz and Ira Kemelmacher-Shlizerman from the University of Washington, released a paper in July 2017 describing how they created the fake Obama.

They developed an algorithm that took audio and transposed it on to a 3D model of the president’s face.

The task was completed by a neural network, using 14 hours of Obama speeches and layering that data on top of a basic mouth shape.

Dr Suwajanakorn acknowledged that “fake videos can do a lot of damage” and needed an ethical framework.

“The reaction to our work was quite mixed. People, such as graphic designers, thought it was a great tool. But it was also very scary for other people,” he told the BBC.

Political crisis

It could offer history students the chance to meet and interview Holocaust victims, he said. Another example would be to let people create avatars of dead relatives.

Experts remain concerned that the technology could create new types of propaganda and false reports.

“Fake news tends to spread faster than real news as it is both novel and confirms existing biases,” said Dr Bernie Hogan, a senior research fellow at the Oxford Internet Institute.

“Seeing someone make fake news with real voices and faces, as seen in the recent issue about deepfakes, will likely lead to a political crisis with associated calls to regulate the technology.”

Deepfakes refers to the recent controversy over an easy-to-use software tool that scans photographs and then uses them to substitute one person’s features with another. It has been used to create hundreds of pornographic video clips featuring celebrities’ faces.

Dr Suwajanakorn said that while fake videos were a new phenomenon, it was relatively easy to detect forgeries.

“Fake videos are easier to verify that fake photos because it is hard to make all the frames in video perfect,” he told the BBC.

“Teeth and tongues are hard to model and could take another decade,” he added.

The researcher also questioned whether it made sense for fake news creators to make complex videos “when they can just write fake stories”.

Outlaw or ignore? How Asia is fighting ‘fake news’

Everybody is talking about it: fake news.

President Trump decries it every time he sees a critical article, the Pope has condemned it, governments are fretting about its influence, holding parliamentary hearings.

And now Malaysia has passed a law criminalising it, with a penalty of up to six years in jail. Yet no-one has defined what it is.

The term first came to prominence during the 2016 US presidential election campaign. But the problem of deliberately falsified news articles, masquerading as properly-researched journalism, goes back centuries.

However, the Malaysian government’s definition in the recently-passed law is far more sweeping than that.

It has criminalised the dissemination of “any news, information, data and reports, which is or are wholly or partly false, whether in the form of features, visuals or audio recordings or in any other form capable of suggesting words or ideas”.

Human rights groups have been quick to point out that this could be used against anyone who makes an error in their reporting or social media posts.

Moreover at least one member of the government has already stated that, when it comes to articles critical of Prime Minister Najib Razak, especially over the notorious 1MDB scandal, where billions of dollars of a government-run investment board are alleged to have been misappropriated, any information not verified as true by the government will be viewed as fake news.

The fact that this law has been rushed through right before what is likely to be a hard-fought general election has raised suspicions that its real purpose is to intimidate government critics.

It is not clear anyway that Malaysia has a serious fake news problem.

In a response to the concerns expressed about the new law, the communications and multimedia minister Salleh Said Keruak highlighted the foreign media’s failure to get the sometimes complicated string of official titles for high-ranking Malaysians right – irritating, yes, but hardly a threat to national security.

The article goes on to excoriate mainstream media which have published negative pieces about Mr Najib, calling them fake news, and thus rather confirming suspicions that the law is aimed at them, rather than the manipulation of social media opinion through fraudulent Facebook accounts and automated Twitter bots.

‘Better safe than sorry’

Singapore is the other country which has raised the alarm over fake news, holding 50 hours of parliamentary hearings.

Facebook’s policy director Simon Milner was publicly dressed down by the law and home affairs minister K Shanmugam over his failure to acknowledge the full extent of data taken by the data analysis company Cambridge Analytica when he testified to the British parliament earlier this year.

Academics speaking at the Singapore hearings presented an alarming scenario of disinformation campaigns launched by foreign actors bent on attacking the island state, of cyber armies in neighbouring Malaysia and Singapore working as proxies for other countries in undermining national security.

It also gave Singapore academics and officials an opportunity to snipe at the US belief in free expression, the “marketplace of ideas”, which had allowed the abuse of personal data on Facebook to take place, in contrast to Singapore’s “better safe than sorry” belief in a more tightly regulated society.

  • The (almost) complete history of ‘fake news’
  • ‘Fake news’: What’s the best way to tame the beast?
  • The city getting rich from fake news

    But the actual examples of fake news which have come up during this national debate have mostly been prosaic; a hoax photo showing a collapsed roof at a housing complex, which sent officials rushing unnecessarily to the scene; and an erroneous report of a collision between two trains on the light rail transit line.

    Irritating and worrying for some, for a while, but hardly likely to bring Singapore society to its knees. In any case both Singapore and Malaysia already have plenty of laws capable of penalising false, inflammatory or defamatory comment.

    A poisonous tide

    In the country where social media misinformation has had the most devastating impact, by contrast, there is no clamour for a fake news law.

    Myanmar too has a raft of existing harsh laws sweeping enough to stifle any reporting deemed a threat to the state or society, laws which all too often have been used to jail journalists.

    But these laws have been unable to prevent a poisonous tide of hate speech on social media, which has helped ignite anti-Muslim sentiment.

    • UN: Facebook has turned into a beast in Myanmar
    • Myanmar conflict: Fake photos inflame tension
    • Rohingya crisis: Suu Kyi says ‘fake news helping terrorists’

      Myanmar famously leapt from being a society largely without even old-fashioned telephone lines, to one with more than 40 million mobile phone accounts, in just three years.

      Seventeen million people have Facebook accounts, and as in so much of Asia, this is how most Burmese send messages and get their news.

      Most don’t bother with email accounts. This has coincided with the end of strict military censorship, and the emergence in the mainly Buddhist population of a primeval fear of the small Muslim minority.

      It has been all too easy to find cartoons and doctored photographs on Facebook which depict Muslims in a sinister and derogatory way. Worse still, large numbers of posts about Muslims are completely false, with photographs purporting to show atrocities against Buddhists by Muslims which are from a completely different part of the world.

      The government has done nothing to stem this tide of disinformation, at times appearing to encourage it.

      For example the Facebook page of the Myanmar armed forces still has on it a gruesome photograph with a caption stating that the dismembered bodies of infants, being dragged by apparently Muslim men, are Rakhine Buddhists killed by Rohingya militants in 1942.

      In fact the photograph is from the Bangladesh independence war in 1971.

      When journalists, myself among them, were given photographs in September 2017 during a government-run tour of Rakhine state, supposedly showing Muslims burning down their own homes, backing the assertion by officials that this was the cause of the destruction of Rohingya villages, we were quickly able to ascertain that the perpetrators in the photos were actually displaced Hindus dressed up as Muslims.

      Yet the government spokesman posted one of the photos on his Twitter feed proclaiming “It’s Truth”, although he later removed it. I was told in all seriousness in one Rakhine village that Muslims used to cut up Buddhists and cook them with their beef stew.

      These kinds of stories are circulating unchallenged in Myanmar, creating a tide of fear and hate which then intimidates anyone trying to advocate a more tolerant approach into silence.

      The UN Special Rapporteur to Myanmar Yanghee Lee, who has herself been subjected to vicious online abuse for her focus on human rights in Rakhine, and has now been banned from entering the country, has described Facebook there as “a beast”.

      Facebook says it takes the problem of hate speech very seriously, but has yet to stop the site being used to stir up sectarian conflict.

      The social media game changer

      The other country where social media has had a profound impact is the Philippines, where critics of President Duterte have accused his supporters of “weaponising” Facebook and Twitter to twist public opinion and silence dissent.

      Filipinos are among the heaviest users of Facebook in Asia, with more than one third of the population visiting the social media site regularly.

      This is has made it a potentially game-changing arena for political actors who know how best to use it, in a country which has long had a lively and competitive traditional media.

      Long before the 2016 election which propelled Rodrigo Duterte, a late candidate with outsider status, to the presidency, the internet was already being exploited by public relations experts promoting products and opinions with so-called “click factories”, where thousands of low-paid workers raised the clicks for specific websites, and companies openly offering hundreds of fake Facebook or Twitter accounts in support of the online profile of clients.

      After announcing his candidacy in November 2015 Rodrigo Duterte hired social media experts to craft a strategy which outflanked the usual dependence on endorsement from mainstream newspapers and television channels.

      It worked brilliantly, tapping into a hitherto unarticulated yearning for change among many Filipinos.

      But researchers have also detected what they believe is the use of automated bots and fake Facebook accounts to amplify the pro-Duterte message, something the president’s team has denied.

      The online news site Rappler published a detailed report about this in October 2016, enraging Mr Duterte’s supporters, and, it believes, prompting the ruling in January this year by the Philippines Securities and Exchange Commission that the site is illegally owned by foreign investors, a claim first made by the president last year.

      Rappler also highlighted the way Facebook’s algorithms could be “gamed” to ensure certain content dominates users’ newsfeeds.

      Leaving aside the allegations of social media manipulation, what President Duterte’s supporters have succeeded in doing is using Facebook and other sites to wage a war of words against his critics, or anyone publishing unfavourable reports.

      I experienced this in September 2016 after the BBC published a report on Mr Duterte’s campaign against drugs dealers and users which has resulted in thousands of police and extrajudicial killings.

      • Philippines’ Duterte admits personally killing suspects
      • The human scars of Philippines drug war
      • Why Rappler is raising Philippine press freedom fears

        I received a flood of hostile messages, and a few death threats on my Facebook page, and the BBC complaints site was swamped with almost identical protests over “erroneous and biased” reporting. Maria Ressa, the CEO and founder of Rappler, was at one point getting 90 hate messages an hour.

        In part this militant response has been shaped by Mr Duterte’s own depiction of his presidency, more as an existential struggle to save his country than just another administration.

        Mr Duterte uses emotive and bellicose language to describe his mission, openly threatening to kill those who stand in his way, including journalists, and suggesting he may in turn be killed, by unnamed enemies.

        Having successfully motivated those who voted for him into believing he could be a one-man saviour for the many ailments afflicting the Philippines, he, like President Trump in the US, has also instilled in his supporters a deep mistrust of traditional mainstream news sources, as controlled by powerful vested interests set on ensuring the failure of his presidency – “presstitutes”, in their preferred term.

        Nowhere has the political climate been more polarised than in Myanmar and the Philippines.

        Yet hearings at the Philippines Senate concluded that a specific fake news law was unnecessary, and possibly counterproductive.

        Clarissa David, a professor of mass communications at the University of the Philippines, testified to the Senate about the dangers of an information environment she described as “polluted”, with no one sure any more what is real and reliable, and what is fake.

        But she warned against easy definitions of fake news. And trying to outlaw it, she argued, is not worth the inevitable cost there will be for media freedom.

        Hers is an argument which was made, but lost, in Malaysia.

Malaysia seeks 10 year-jail terms for ‘fake news’

Malaysia’s government has proposed new legislation to combat “fake news”, with offenders facing up to 10 years in prison.

Under the Anti-Fake News Bill, those convicted of disseminating false content would be jailed or fined up to RM500,000 ($128,000; £90,400), or both.

The bill was tabled in parliament on Monday, ahead of a national election that is expected within weeks.

Critics have called the bill an attempt by the authorities to stifle dissent.

According to the bill, the term “fake news” is defined as “news, information, data and reports which is or are wholly or partly false”.

An offender would be anyone who “by any means, knowingly creates, offers, publishes, prints, distributes, circulates or disseminates any fake news or publication containing fake news”.

Blogs, public forums and social media accounts are also covered by the bill.

It applies to anyone inside or outside Malaysia, as long as the “fake news” published concerns the country or those in it, which means foreigners can technically be sentenced in absentia.

True or false?

The move has been described as an attempt to silence opposition ahead of the general election. It must be held by August, but is widely expected within the next few weeks.

“The bill is 100% intended to muffle dissent… the punishment is extremely high and what amounts to fake news has been loosely defined,” Eric Paulsen, co-founder of Malaysian human rights group Lawyers for Liberty, told the BBC.

“It also looks like they are rushing through the bill… before the elections. It’s likely that it will be passed.”

He added that the bill could affect reporting stories such as the scandal surrounding 1MDB, Malaysia’s state development fund. It is alleged that hundreds of millions of dollars were “misappropriated” from the fund.

  • What is 1MDB?

    The Wall Street Journal (WSJ) had in 2015 reported it had seen a paper trail that allegedly traced close to $700m from the fund to Malaysian Prime Minister Najib Razak’s personal bank accounts.

    Mr Najib has denied taking any money from 1MDB.

    Just days ago, a Malaysian minister said any news relating to 1MDB that was not confirmed by the government was false.

    “If this is [the government’s] benchmark, then we will have a situation where only the government can determine if things are true or false,” said Mr Paulsen.

Malaysia seeks 10 year-jail terms for ‘fake news’

Malaysia’s government has proposed new legislation to combat “fake news”, with offenders facing up to 10 years in prison.

Under the Anti-Fake News Bill, those convicted of disseminating false content would be jailed or fined up to RM500,000 ($128,000; £90,400), or both.

The bill was tabled in parliament on Monday, ahead of a national election that is expected within weeks.

Critics have called the bill an attempt by the authorities to stifle dissent.

According to the bill, the term “fake news” is defined as “news, information, data and reports which is or are wholly or partly false”.

An offender would be anyone who “by any means, knowingly creates, offers, publishes, prints, distributes, circulates or disseminates any fake news or publication containing fake news”.

Blogs, public forums and social media accounts are also covered by the bill.

It applies to anyone inside or outside Malaysia, as long as the “fake news” published concerns the country or those in it, which means foreigners can technically be sentenced in absentia.

True or false?

The move has been described as an attempt to silence opposition ahead of the general election. It must be held by August, but is widely expected within the next few weeks.

“The bill is 100% intended to muffle dissent… the punishment is extremely high and what amounts to fake news has been loosely defined,” Eric Paulsen, co-founder of Malaysian human rights group Lawyers for Liberty, told the BBC.

“It also looks like they are rushing through the bill… before the elections. It’s likely that it will be passed.”

He added that the bill could affect reporting stories such as the scandal surrounding 1MDB, Malaysia’s state development fund. It is alleged that hundreds of millions of dollars were “misappropriated” from the fund.

  • What is 1MDB?

    The Wall Street Journal (WSJ) had in 2015 reported it had seen a paper trail that allegedly traced close to $700m from the fund to Malaysian Prime Minister Najib Razak’s personal bank accounts.

    Mr Najib has denied taking any money from 1MDB.

    Just days ago, a Malaysian minister said any news relating to 1MDB that was not confirmed by the government was false.

    “If this is [the government’s] benchmark, then we will have a situation where only the government can determine if things are true or false,” said Mr Paulsen.

Reddit admits hosting Russian propaganda

Reddit has become the latest social-media platform to admit that Russian propaganda was used on its site during the 2016 US presidential election.

It follows leaks from news site The Daily Beast showing a Russian troll farm active on the website.

Co-founder Steve Huffman said that it had removed “a few hundred accounts” suspected of being of Russian origin.

In a blogpost, he said “indirect propaganda”, which was more complex to spot and stop, was the biggest issue.

“For example, the Twitter account @TEN_GOP is now known to be run by a Russian agent. Its tweets were amplified by thousands of Reddit users, and sadly, from everything we can tell, these users are mostly American and appear to be unwittingly promoting Russian propaganda.”

Conspiracy theories

Mr Huffman added: “I believe the biggest risk we face as Americans is our own ability to discern reality from nonsense, and this is a burden we all bear.

“I wish there was a solution as simple as banning all propaganda, but it’s not that easy. Between truth and fiction are a thousand shades of grey.

“It’s up to all of us—Redditors, citizens, journalists—to work through these issues.”

The @TEN_GOP account appeared to be run by Republicans in Tennessee. It tweeted a mix of pro-Trump content and conspiracy theories, as well as more obvious fake news stories.

The Daily Beast investigation suggested no outright support of any particular candidate or viewpoint and concluded that Russia’s aim was to provoke and divide Americans on the internet and, as a result, in the physical world too.

Social media ‘weapon’

Social media platforms are under increased scrutiny from the US Congress over the issue of Russian meddling in the 2016 election.

Facebook has given the Senate Intelligence Committee thousands of ads believed to have been purchased by Russian agents.

The Washington Post reported that Reddit was now likely to be questioned over its involvement in the “weaponisation of social media” during the election.

Special counsel Robert Mueller has charged 13 Russians with interfering in the US election, all of whom are linked to troll farm the Internet Research Agency.

Meanwhile, pressure is mounting on Reddit to clean up the content on its platform.

In February, it banned a group that was generating fake porn – imagery and videos that superimpose a person’s face over an explicit photo or video without permission.

This week, it emerged that another subreddit was sharing images of dead babies and animals being harmed.

Mr Huffman said the company was aware of the group, which currently has nearly 19,000 subscribers, and that the community was “under review”.