Elections this week in Ireland and the Netherlands were disrupted by AI deepfakes as the post-truth future that experts have long warned about came one step closer.
Newly elected Irish President Catherine Connolly survived a doctored video showing her supposed withdrawal from the election on the eve of voting, while Dutch firebrand Geert Wilders was forced to apologise for a fabricated video distributed by two of his party’s MPs depicting centre-left opponent Frans Timmermans being arrested.
Since deepfakes first emerged in 2017 as “incel-produced nonconsensual porn”, concerns have “snowballed into panic” when their political consequences became apparent, said The Guardian. AI “slopaganda” is here to stay and promises to influence our lives “for better or for worse”.
What did the commentators say?
“AI-generated content is being deployed to sway minds,” said Politico. Fake content in the recent Irish and Dutch elections “exposed significant gaps” in structural efforts to ensure accuracy and to prevent the exploitation of the electorate.
Some voters “may have been surprised” to see Connolly’s name on the ballot sheet after a video appeared that said: “I announce the withdrawal of my candidacy and the ending of my campaign”. It included convincing material with two well-known TV presenters discussing the implications of the removal on a fake bulletin on national broadcaster RTÉ.
In the Netherlands, AI fakes “overshadowed” what was a pivotal election, where the “plethora” of minority parties means “finding a majority will not be easy”, said Channel 4 News. The landscape is ripe for exploiting division. Voters are “tired of the constant mudslinging” and “tit-for-tat” debates.
Only a week before, the Dutch data regulator had expressly warned voters against using AI chatbots to inform their decision, saying online platforms issue “unreliable advice and push them towards two major parties on opposite ends of the political spectrum”, said Reuters.
This is not just a Dutch problem. Advances in technology have made it easier than ever for individuals to create election-altering fake videos, said Abbas Yazdinejad and Jude Kong on The Conversation. The rapidly evolving landscape is bursting with videos that are “shockingly simple to create and near‑impossible to detect”. The implications are stark and require urgent intervention. The “myriad” disinformation threats could “erode public trust” and spell the end of conventional political election contests.
We’re at an uncomfortable crossroads. With electorates becoming increasingly drawn to short-form video content, voters are caught between online platforms that are “not foolproof” and accelerating technology that “continues to improve”, said EU Observer.
Deepfakes by “bad actors, political parties and candidates themselves” have become a “feature” of global politics. There has been plenty of commentary warning voters of deepfake imagery, but only recently are we seeing it slip consistently into election campaigns and criticism.
What next?
The Irish presidential election may be “small potatoes” compared to other elections around the world, said Futurism. However, the lack of regulation twinned with greater reliance on AI to sift through information in this election emits a “glaring signal” to Meta and other social media companies that electorates are “incredibly vulnerable” to “malicious interference”.
Going forward, legal particulars need to become more defined and easier to implement, said Politico. Though there is no legal framework on digital likeness rights that is EU-enforced, there is an EU-specific law regarding “labelling” artificial intelligence, which could be a “big part of the response”.
Next month, Brussels is due to put forward an initiative concerned with “upholding the fairness and integrity of election campaigns against foreign manipulation and interference”. However, this is not expected to contain “any binding legal requirements”.
AI ‘slopaganda’ is becoming a ‘feature’ of modern elections
