Shortly before Prime Minister Justin Trudeau called the 2021 federal election, his government rolled out a proposal that would hold online service providers accountable for a variety of internet maladies. The framework aims to eradicate child pornography, child sexual exploitation, terrorist, violent and hateful content.
Cleaning up the internet is necessary, but the Liberal government’s idea introduces severe conceptual and logistical concerns which negate its good intentions.
In the proposal, Online Communication Service Providers (OCSPs) are expected to address all flagged content within 24 hours of flagging, either by removing content altogether or specifying the flag does not meet the criteria for harmful content.
Upon closer examination, however, the proposal does not outline how OCSPs are supposed to accommodate the sheer volume of flags and reports. This is problematic considering how easy it is to flag something online. Answering every flag would be nearly impossible, given this can be done with the (literal) click of a button.
Realistically, the solution involves automation. But automated content removal is far from perfect and its application in complex socio-political environments is fraught with human and tech-based issues.
Yet, the proposal insists that a tribunal may fine an amount up to three per cent of the company’s gross global revenue or $10 million, whichever is higher, in the event of rules violations under the proposal. For Facebook, three per cent of gross global revenue is $870 million USD, a hefty sum to part with.
Moreover, attempting to strongarm online service giants may lead to more pushback than expected.
Earlier this year in Australia, for example, a proposed law to establish a “fairer” process for Facebook and Google to publish news stories provoked them to block this content in the country. Google even threatened to withdraw its primary search engine from the country.
The news blackout’s eventual restoration is telling: the Australian government amended its decision and Facebook retained its previous news policy.
The term “amendment” may help appearances, but the outcome is clear—the Australian government was bullied out of its proposal.
A similar response to the Canadian regulatory framework is, at least, feasible.
Still, the issues with this proposal extend beyond logistical concerns or hurting relationships with online media giants.
Citizen Lab, an interdisciplinary lab based out of the University of Toronto, submitted a letter to the Canadian federal government outlining the potential harm associated with this proposal.
The document points out that while technology-facilitated violence is a very real issue, it unequally impacts minorities. Citizen Lab references research indicating that “internet filtering and content monitoring technologies often result in disproportionate censorship and surveillance of historically marginalized individuals and communities.”
Ironically, the same content-filtering measures aimed to protect marginalized groups from hate speech will serve to disproportionately censor those groups.
The proposal advances “an aggressive, algorithmic, and punitive regime for content removal it proposes, without any substantive equality considerations or clear safeguards against abuse of process,” the letter says.
Citing a study titled “The Risk of Racial Bias in Hate Speech Detection,” the letter points out “hate speech detection” algorithms often demonstrate anti-Black bias. Such effects will only exacerbate existing issues revolving around racism and discrimination.
Among a variety of concerns, Citizen Lab also expresses fear over new Canadian Security Intelligence Service (CSIS) powers that the proposal introduces.
If the CSIS Act is modified as proposed, the intelligence agency could collect basic information from any subscriber to an OCSP without satisfying specific warrant requirements that are currently in place.
The proposal raises serious questions about the nature of online policing and widespread surveillance. “The fact that the government’s proposals would explicitly deputize technology companies in the surveillance and policing of their users on behalf of Canadian law enforcement and intelligence agencies is all the more disturbing,” reads the letter.
Creating a legal obligation to monitor all users and moderate all content while further enabling CSIS to mysteriously investigate behind the scenes screams privacy concern.
“Individuals who have their information disclosed to CSIS may never be able to contest such disclosures, or contest how that information is ultimately used,” the Citizen Lab letter says.
Scathingly, the letter concludes with a section titled “Rewrite the Proposal From the Ground Up.” Considering the broad range of issues, the government should cut their losses and do just that.
Featured graphic by Sara Mizannojehdehi.
