Digital Deception in Warfare: The Perils of AI-Generated Deepfakes in the Gaza Conflict

33

In the recent Gaza conflict, amidst the stark images of destruction, certain pictures have been particularly harrowing: seemingly injured, forsaken babies. These viral images, however, are not real. They are AI-generated deepfakes, distinguishable upon close inspection by their slightly unnatural features like oddly bending fingers or unusually shimmering eyes. Yet, the anger and horror they incite are undeniably genuine.

These deepfakes serve as a grim testament to AI’s growing role in fabricating realistic images of violence and devastation. With the proliferation of digitally altered content on social media, these images have been falsely attributing blame or fabricating non-existent atrocities since the conflict’s onset.

The rapid advancements in AI technology, coupled with minimal regulation, underscore its potential misuse as a tool of misinformation in warfare and other significant global events. Jean-Claude Goldenstein, CEO of CREOpoint, highlights the escalating threat of AI in generating misleading pictures, videos, and audio. CREOpoint has compiled a database of the most viral deepfakes from the Gaza situation.

Not only are old photos from different conflicts being misrepresented as current, but new images are also being fabricated entirely using generative AI. This includes a viral image of a crying baby amidst bombing ruins. Such AI-generated content, including fabricated videos of missile strikes or tanks in destroyed areas, is often emotionally charged, featuring infants and families to evoke stronger reactions.

Imran Ahmed, CEO of the Center for Countering Digital Hate, points out that whether these images are real or fabricated, their emotional impact on viewers remains the same. The more disturbing the image, the more likely it is to be shared, further spreading misinformation.

The use of deceptive AI content is not limited to the Gaza conflict. For instance, altered videos from the 2022 Russia-Ukraine conflict made false claims, like a video purporting to show Ukrainian President Volodymyr Zelenskyy urging surrender.

As major elections approach in countries like the U.S., India, Pakistan, Ukraine, Taiwan, Indonesia, and Mexico, concerns are rising about the misuse of AI for disinformation. This has prompted bipartisan concern in the U.S., with lawmakers like Rep. Gerry Connolly emphasizing the need for AI tools to counter such threats.

Efforts are underway globally to develop technologies capable of identifying deepfakes, authenticating image origins, and assessing the accuracy of AI-generated content. Maria Amelie, co-founder of Factiverse, mentions their AI program designed to detect inaccuracies or bias in AI-generated content, highlighting its importance for educators, journalists, and financial analysts.

However, according to David Doermann, a computer scientist and former DARPA project lead, staying ahead of AI-generated disinformation requires more than just technological solutions. It necessitates a combination of improved regulations, industry standards, and digital literacy initiatives, as those creating AI falsehoods are often one step ahead in masking their traces.

Frequently Asked Questions (FAQs) about AI-generated deepfakes

More about AI-generated deepfakes

• Understanding Deepfakes
• The Gaza Conflict and AI Misinformation
• Challenges of AI in Journalism
• CREOpoint and AI Verification
• Center for Countering Digital Hate
• Future of AI in Misinformation Campaigns
• Technology Against Deepfake Threats
• Factiverse: Combating AI Biases
• AI Disinformation in Global Politics
• Digital Literacy in the Age of Deepfakes