My work at The Disinformation Project in New Zealand has since the last quarter of 2022 led me to conduct significant, and sustained research into the adoption, adaptation, and application of generative AI models, and methods in the domain of information integrity, truth decay, and disinformation. Put another way, I’ve experimented with hundreds of generative AI products, platform, apps, and services to find out how, and to what degree, they can create or manipulate content that can impact things like electoral integrity, referendums, political campaigns, civil society activism, political communications, public consultations, and judicial processes including impact on issues like the chain of custody, and permissible evidence in court proceedings. In January 2023, in a confidential brief, I identified ten key threats to New Zealand’s democracy as a consequence of generative AI’s vaulting progress, early adoption by the far-right, and rapid adaption.
This multi-domain (graphics, video, text, audio) research now, somewhat strangely, intersects now with the study of Sri Lanka’s retrograde Online Safety Act (OSA). To this end, I spent some time this weekend working through some hypothetical scenarios featuring the combined weaponisation of generative AI, along with specific clauses in the OSA in order to silence specific individuals, institutions, and ideas.
It’s not looking good for the country.
Generative AI, and the OSA present, entirely independent of each other, significant issues to Sri Lanka’s fragile democratic fabric. A Zoom lecture on generative AI in July 2023 was the first I know of in the country, in Sinhala, to outline some of the risks, and threats to social cohesion, fundamental rights, electoral integrity etc especially in the context of a post-aragalaya democratic deficit. While in New Zealand, the Chief Science Advisor, and others now look at the pitfalls, and potential of generative AI in the context of democratic processes, and pillars, no such process or sense of purpose is present in Sri Lanka.
The OSA is in a sense simpler to present. It kills democracy, and impacts everyone from the advertising industry to trade unions. The Act is irredeemably, indubitably, and enduringly detrimental to country. Amendments to the Act, proposed by a misguided few as somehow ‘salvaging’ the Act, are terribly misguided, and without any merit. The Act is a party blind debacle, and will inexorably, and irrevocably harm democracy the longer it is in our statute books.
Combine generative AI’s advances, and adversarial adaptation (in Sri Lanka) with the OSA’s draconian substance, and you have an incredibly nightmarish combination that’s hard for many to grasp. This is why I spent the weekend realising content around some future scenarios, which I shared with a select few – and to their abject horror. I can’t, for obvious reasons, share in public these future scenarios (why give ideas to those who are already hell bent on killing democracy?). I can however share the no one I shared the content with remotely understood how the OSA could be weaponised using generative AI, till it was made evident to them. The individuals I shared the content with aren’t unintelligent, uninformed, or disconnected from Sri Lanka’s social, political, and economic pulse. One rated the believability of something I shared as 8/10, which horrified the person because it was related to the generation of content which appeared to be from a person who was intimately known, for decades. In other instances, content which appeared to be from leading political, business, and media figures helped communicate how generative content in Sinhala, Tamil, and in various forms ranging from text to audio could easily shape, and shift perceptions, and reactions of specific audiences, communities, and markets.
Suffice to say that I comprehensively ruined the weekend for a few who now know, more than the majority in the country, how the OSA can be used this year, and for as long as it remains in our statute books, against those who critique power, and those who wield it in politics, and other areas like industry, the military, non-state actors, and even around foreign state actors who have already architected influence operations in Sri Lanka.
From January this year onwards, I’ve found myself leading conversations around the OSA amongst various industry leaders, trade unions, academics, lawyers, journalists, civil society, youth, activists, ambassadors, diplomats, and representatives of social media companies who have, in what’s still hard for me to grasp, little to no idea about the Act’s substance or appreciate it in the context of Sri Lanka’s enduring investments on, and interest in surveillance, the suppression of dissent, suffocation of rights, and subjugation of democracy to Executive fiat.
Generative AI’s vaulting developments render all this much worse. Already, there are regulatory frameworks in the US, for example, around the use of generative AI in political campaigns. Instead of this timely, enduringly important, and wide-ranging discussion, Sri Lanka’s OSA renders a dystopian future which, incredibly, and ironically, expands (instead of constricts) socio-political theatres for the weaponisation of generative AI, disinformation, and truth decay in general.
I have found that until individuals are shown what’s probable, and possible, they remain oblivious to how dangerous the OSA is. And it is this very ignorance that our politicians continue to benefit from.
Sanjana Hattotuwa 