VANCOUVER—Facebook Canada’s head of public policy admits the social media giant needs to “do more” about online hate.
During a one-on-one interview with Star Vancouver, Kevin Chan discussed the company’s plans to combat foreign interference, the spread of misinformation and online hate in a murky, borderless digital world.
The online platform made a stop in Vancouver Tuesday as part of its cross-Canada tour to encourage civic engagement ahead of Canada’s federal election. It also launched its new transparency tools in Canada on the same day to comply with C-76, Canada’s Election Modernization Act, which came into effect last year.
The law is aimed at stopping bad actors — whether foreign or domestic — from interfering with Canada’s upcoming election through advertising.
Tuesday’s launch focused on tools for political advertising, which include: steps to confirm that an individual or group buying a political ad is real and based in Canada; making who paid for the ads visible as well as giving users an option to view its demographic reach; and creating a digital library that will hold all political ads for seven years.
However, unpaid content or content outside the realm of political advertising — such as misinformation or online hate — is still a “complicated” challenge.
Here’s a Q&A with Chan about the changes, misinformation, election interference and online hate. It has been edited for clarity and length.
How does Facebook see the line between personal responsibility of the user and the platform’s responsibility to combat foreign interference, misinformation and online hate?
We obviously have a responsibility to do our part. You can see that in the way that we are trying to put structure and process around political advertising. We have community standards, which are a global policy that governs what can stay up and what should go down on the platform. It’s very detailed. This is not something that can just be taken care of by an online platform on its own. There is how to look broadly at the ecosystems about digital literacy. And then the third piece is the role of government. We have community standards, we are wrestling with these difficult challenges. But governments themselves also have to then weigh out what baseline standards are. So what is permissible and not permissible online? These are questions that if the parliaments of the world wish to regulate, we would welcome that. It would provide clarity and baseline standards for everything on the internet. This is something that we have to work together on. Everyone. Civil society, governments and online platforms.
Fake or manipulated videos, online hate and Islamophobia are prevalent even with the community standards. How is it some things stay up while others don’t?
These are fascinating issues but also important issues that we have to work hard to get right. When you’re talking about people’s speech online, there’s going to be a lot of grey zones. There are some black and white cases: Pornography can’t be allowed on our platform, so we take that down. But in fact if you look at our community standards, we talk about nudity and nudity can’t be on the platform. How about images that are public interest? You know the iconic photo from Vietnam, which was taken amid the war. That one, on its face is nudity and in fact it’s child nudity. On one level, the community standards would remove that. And it did, initially when this first came up. We learned from that there’s probably some category of content that is newsworthy that has value beyond the post itself, like historical value. I think your question gets at this point, which is that it’s not a good idea, we don’t think, for Facebook to be making all of these decisions on their own.
And so in addition to evolving our community standards, working with experts to better reflect them, we’re also looking at building an independent oversight board by the end of 2019. This board would be independent from Facebook and it would be a final board of appeal for these kinds of really tough and precedent setting cases.
Research has linked foreign interference to online hate in the sense that foreign actors can influence elections by using divisive rhetoric on wedge issues on Facebook. What is the company’s long-term plan to deal with online hate?
I would draw a big distinction between foreign interference, where there’s a foreign entity or nation state that’s trying to interfere with democratic elections and harmful content, which could, quite frankly, be foreign or domestic. We can have people within the country that say things that you may disagree with or find hateful. They are different.
On the harmful content side, we do have community standards. They’re very clear. That’s quite advanced, it’s not just a baseline of what is considered hate speech in Canada, we go way higher than that. Any hateful speech has no place on Facebook, in the sense that we would take down bullying content, we would take down incitement of violence, we would take down white nationalist content or things targeting particular protected groups.
Much of that content remains up on Facebook despite the community standards. Can you speak to that?
There are instances where people will have views and those views may be offensive or difficult or challenging for others, but if they are not in violation of the community standards, we do want to give people as much space to express themselves as possible. That goes to the heart of what we mean by freedom of expression. It goes to the heart of what we mean by a free and open internet. There are 24 million Canadians on Facebook. Not everyone is going to agree. And even if one per cent of that group has particular views that are outside of what others may consider or the majority may consider to be polite or appropriate thing to say, that’s still a sizable amount of people and those people should be given that ability to express themselves. So long as it’s not a violation of the public laws of the land, and it’s not a violation of community standards.
A lot of people post every day on Facebook. And so the question is, how do we address (community standards) in a way that we can scale? We are working on automated systems to try to do that better. Humans need to review this content for context. Part of it, for flagging purposes, should come from artificial intelligence that we’re developing.
In our latest May transparency report, there are stats specifically on hate speech. When we started on this journey about a year and a half ago, we were automatically flagging in the system about 24 per cent of hate speech. In the last year and half or so, we’ve gotten that upwards of 60 per cent. That content is getting automatically flagged in our system. We want to get it even further up. This is an ongoing process. These technologies are very new. They’re not perfect and hate speech is, in some ways, much harder to get at because context is so important.
We want to go further but we’re not there, as opposed to terrorist content. If you go back about five years ago, there was a lively debate about what if we find terrorist content on Facebook? Right now we automatically flag and remove about 97 per cent of the content, most within seconds of upload. We want to apply the same playbook to terrorist content that we did more or less successfully to removing hate speech.
But hate speech is a lot greyer, it’s a lot harder to recognize. That one is going to be a much harder and complicated challenge. We are making a lot of progress and we obviously need to do more.
Facebook is laughing at Canada’s toothless privacy laws