AI in election campaigning: Harassment surges amid deepfake content

Analysts advise against sharing information or videos without verification

Artificial Intelligence (AI) has introduced a new dimension to social media campaigning in this year’s election.

Its misuse, however, is placing significant strain on the parties’ campaign efforts.

As polling day approaches, the proliferation of fabricated videos, images, and audio clips is driving a surge in misinformation, confusion, and online harassment against candidates and political parties.

Rival political parties, analysts, and election officials report that fake content is being circulated across social media platforms -- specifically Facebook, YouTube, and TikTok -- to target opponents ahead of the parliamentary election.

This is creating additional pressure on the electoral environment, voter confidence, and campaign management.

The fraudulent material includes "deepfake" videos, cloned voice recordings, clips mimicking legitimate news bulletins, and AI-edited posters and graphics.

Victims and experts said in many instances, the content is crafted so convincingly that it is difficult to distinguish from authentic news or genuine statements.

Consequently, voters are being misled, sparking unintended reactions.

According to analysts, AI-generated disinformation does not merely confuse voters; it possesses the potential to destabilise the political climate and undermine law and order.

They have urged the public not to share any information or video without prior verification, advising citizens to rely on the Election Commission and credible news outlets for election-related updates.

Political campaign teams and observers suggest that AI-generated fake content is no longer an isolated occurrence.

One side is intentionally employing these tactics to embarrass opponents, establishing a "new reality" in the electoral battle.

Many clips feature party leaders "uttering" statements they never actually made.

Techniques include lip-syncing videos, dubbing over audio with cloned voices, or repurposing old footage with new, fabricated dialogue.

There are also allegations of "paid and targeted" smear campaigns designed to mimic organic grassroots support.

Moreover, campaign officials claim that footage of violence from previous years or different countries is being recirculated and framed as local political incidents.

This raises serious concerns about misinformation and heightened tensions over public safety.

The spread of AI-generated screenshots, fake polls, or fabricated reports of arrests is also being flagged as a major factor that could influence voter decision-making.

Analysts also highlight an increased risk of character-assassination and insulting deepfake content targeting female politicians.

[Read more details in Bangla] ভোটের প্রচারে এআই: ভুয়া কনটেন্টে বাড়ছে হয়রানি