Voter suppression threats using @Google @Apple Exposure Notification #GAEN have been dismissed as unrealistic: an adversary could achieve the same effect much more easily using fake news campaigns (e.g. @ronrivest in Forbes). This thread attempts to analyze these arguments (1/12)
2. So ask yourself: what would make you more likely to stay home? A social media post of dubious origin that your polling station is contaminated or a personalized state-sanctioned notification of exposure to #COVID19 ?
3. While disinformation campaigns may be easier to mount, isn't it reasonable to assume that the return (as in the ability to affect a person behavior) is much higher if the attacker messes up with #GAEN ?
4. But, wait, my twitter followers are by no means representative of the population at large. They check their sources and are more likely to believe a #GAEN notification, while the population at large can be more easily fooled. True, maybe
5. But, then, think of this as advertising: we have just given the adversary a strategy to reach a segment of the population that was not so easily reachable by disinformation campaigns. Why wouldn’t the adversary leverage both routes?
6. It might even be reasonable to assume an adoption and compliance bias to #GAEN notifications that would disproportionately affect populations of a particular political leaning (see e.g. the mask wearing adoption bias aligned mostly with personal politics)
7. Given the US electoral system where a small perturbation in voter turnout in specific geographic areas can affect the result, the return on the attack does not even have to be too high to achieve the desired goal
8. See e.g. Gomez et al. [Journal of Politics, 69(3), 2007] where it is argued that lower turnout due to rain in select districts may have had a critical impact on the outcome of at least two U.S. presidential elections
9. Wouldn’t a state sanctioned exposure notification have a similar effect to a rain forecast? Are we giving the adversary the ability to “figuratively” rain on select districts?
10. Additionally, the goal of the adversary may be to simply cast doubt on the electoral process. In a democratic system, public trust in the integrity of the election is as important as whether or not the election's results were changed by some attack
11. Tampering with state-sanctioned public health notification systems has a much larger destabilizing value than launching a fake news campaign on social media, in terms of the potential for compromising citizens' faith in the democratic process
12. Finally for those who think this is still an unrealistic threat: pre-2016 we didn't expect foreign nations to use social media to affect voter turnout. Maybe we should be a bit more proactive now, and prepare for the unexpected side effects of deploying a new technology.
@threadreaderapp unroll
• • •
Missing some Tweet in this thread? You can try to
force a refresh
