Freedom of expression: the holy grail of civil society. Especially in the context of a hyper-digitalised world, freedom of speech online has become one of the most controversial issues worldwide. Just look at former US President Donald Trump for one of the most blatant spread of misinformation online. Trump’s suspension from major social media platforms in the aftermath of Jan 6 ignited an uproar among his supporters, who vigorously protested that he had a right to freedom of speech by presenting his opinions online. There exists a fine line between freedom of expression online, and the blatant spread of misinformation and hate speech. Enter the EU Digital Services Act (DSA).
Image Credit: New York Post
First proposed in Dec 2020 by the European Commission and finally coming into force on 16 Nov 2022, the DSA is a culmination of a single uniform set of obligations and measures aimed at regulating the online environment. The result of months of negotiations between EU member states, the legislation is directed at the rampant abuse and misuse of algorithms by online content platforms - specifically social media sites - that host and amplify the exchange of harmful content such as disinformation, terrorist propaganda and sexual abuse material. By spelling out a comprehensive set of new obligations for these platforms to counter risks online, as well as protecting users’ rights, the DSA seeks to create a unique transparency and accountability framework that aims to set the benchmark for a unified regulatory approach to online platforms across all EU member states. In other words, it consolidates different pieces of existing EU legislation to harmonize the rules applicable to the provision of digital services, thereby avoiding potential conflict between competing national legislation such as Germany’s NetzDG and the revised Avia Law in France.
What was particularly striking about this fancy new legislation is its supposed stance on freedom of expression and speech. The DSA places a specific focus on controlling and limiting the spread of illegal content and products online, aiming to strike a balance between freedom of expression and protecting users from harmful content. Firms are compelled to adopt a more hardline approach to tackling harmful content while steep fines for non-compliance apply. More specifically, online platforms face a monetary penalty of up to 6% of their global turnover if they are assessed to be in violation of DSA rules. When it comes to such fines, Big Tech is well aware that the EU means business. Well, they do not have to look far either: with some US$1.3b in fines for data privacy transgressions levied by various EU data protection authorities under the General Data Protection Regulation (GDPR) in 2021 alone, it is clear that the DSA similarly is no paper tiger.
The chilling effect is obvious - the vagueness of DSA wording serves as a powerful leverage upon companies to remove more content and moderate their platforms in order to avoid falling afoul of the sweeping legislation. Even before the act, companies like Facebook & co engaged in pre-emptive compliance efforts by wording conditions of use to broadly define “malicious content”.
Yet, article 14 of the DSA specifically places the responsibility of determining the legality of user-generated content in the hands of the online platform. The issue herein lies with online platforms: most would rather play it safe given the herculean task for their legal teams to sift through thousands of such content of legally-dubious content on a daily basis. As such, it is natural that companies will choose to employ even stricter preemptive measures to ensure that their moderation is fully compliant with the law, which naturally calls into question the potential repercussions on one’s freedom of expression.
Image Credit: Jimmy Margulies
Another issue that has been raised is the potential for the DSA to be used as a political tool for censorship. The new definition of “illegal content” is broad and includes any information which, in itself or in relation to an activity, is not in compliance with EU or Member States’ laws, irrespective of their subject matter. Essentially, anything that nations deem illegal, which opens up a whole new Pandora’s box about countries’ political systems and governance styles. While Western Europe generally adopts a more liberal style and allows for freedom of expression to a larger extent, Central and Eastern Europe greatly differ in how their governments perceive and permit the freedom of expression within their national borders. Right-leaning Hungary, for example, is very wary of any content that might go against their rhetoric. Viktor Orbán’s regime has repeatedly railed against the “shadow-ban” system employed by social media giants to restrict visibility of so-called undesirable user-generated content - anything from conservative opinions and anti-LGBT discourse to terrorist propaganda. It is not far-fetched to anticipate that governments would define “illegal content” more conservatively and take advantage of the DSA’s broad definitions to censor anything deemed politically detrimental through the use of takedown orders. What guardrails should exist to defend against the weaponization of the DSA in this instance?
But let’s also examine the elephant in the room. The DSA requires online platforms to employ the use of automated filtering systems using AI and machine learning algorithms to detect and remove illegal content. At first glance, this piece of legislation might seem like a smart move: take advantage of technology to automate more processes and in so doing, alleviate the pressure on Big Tech firms to process requests within a matter of hours. However, there are huge flaws with this. One must remember that with our current state of technology, AI is still not the fancy and intelligent software that we envision it to be. With many recent examples pointing to how AI can be manipulated when fed biased data (such as discriminatory practices against African Americans in determining recidivism risk and when determining credit scores in the USA), the message is obvious: AI in its present form is far from ready to tackle this nuanced issue of identifying harmful disinformation versus genuine opinion. When the software becomes error-prone, so too does the likelihood of it flagging lawful content, which might hinder free speech. Worse still, the possibility of AI developing a partisan slant is also very real - after all, previous trials of AI algorithms such as Microsoft’s Tay bot was pulled after a mere 24 hours when Twitter users exploited an unforeseen vulnerability resulting in the bot spewing a variety of homophobic and neo-Nazi slurs. Given the subjective nature of user-generated content, AI can only do so much when it comes to sifting through what remains online and what deserves to be taken down. Human oversight remains the way to go, at least for the time being.
Image Credit: New York Times
While there may be concerns about the DSA’s potential impact on free speech (one need only look to Swedish Member of the European Parliament (MEP) Jessica Stegrud arguments against the DSA), it is important to note that the DSA’s provisions are intended to balance the need to respect users’ rights as well as to protect users from harmful content. For better or for worse, the DSA now enters our history books as a landmark legislation to mass regulate the digital sphere. How it will be written about now depends on how it is implemented and enforced.
References
Article 19. (2021, February 11). At a glance: Does the EU Digital Services Act Protect Freedom of expression? Retrieved from https://www.article19.org/resources/does-the-digital-services-act-protect-freedom-of-expression/
Bodoni, S. (2022, January 18). EU's tough data privacy rules rake in biggest annual fines. Bloomberg.com. Retrieved from https://www.bloomberg.com/news/articles/2022-01-18/eu-s-tough-data-privacy-rules-rake-in-biggest-annual-fines
Digital Services Act: EU's landmark rules for online platforms enter into force. European Commission . (2022, November 16). Retrieved from https://ec.europa.eu/commission/presscorner/detail/en/IP_22_6906
Digital Services Act: Protecting the digital space against the spread of illegal content. Akin Gump Strauss Hauer & Feld LLP. (2022, July 20). Retrieved from https://www.akingump.com/en/news-insights/digital-services-act-protecting-the-digital-space-against-the-spread-of-illegal-content.html
Milmo, D. (2022, April 23). Eu agrees rules to force big tech to rein in illegal content or face huge fines. The Guardian. Retrieved from https://theguardian.com/world/2022/apr/23/eu-agrees-rules-to-force-big-tech-to-rein-in-content-or-face-huge-fines
Stegrud, J. (2022, April 25). The EU's Digital Services Act is undermining free speech. Brussels Report. Retrieved from https://www.brusselsreport.eu/2022/04/25/the-eus-digital-services-act-is-undermining-free-speech/
Comments