Moderating Synthetic Content: the Challenge of Generative AI

Area of Science:

• Digital Ethics
• Artificial Intelligence Safety
• Social Media Governance

Background:

• Generative artificial intelligence (AI) enables the creation of highly convincing fabricated content.
• The proliferation of synthetic misinformation, political propaganda, and non-consensual deepfakes poses a significant threat to the public sphere.
• Malicious applications of generative AI technologies are expected to increase.

Purpose of the Study:

• To evaluate the appropriate policy response for social media platforms to the threat of generative AI-generated content.
• To determine if new, specific policies are required for synthetic media or if existing rules suffice.

Main Methods:

• Analysis of the nature of threats posed by AI-generated content.
• Comparison of AI-generated content threats with existing categories of harmful online content.
• Argumentation based on the principle that the type of harm, not the origin of content, dictates the response.

Main Results:

• The threat posed by AI-generated content is fundamentally similar in kind to that of ordinary harmful content.
• Generative AI amplifies existing problems rather than introducing entirely new categories of harm.
• Existing platform rules, when properly enforced, are adequate to address the challenges.

Conclusions:

• Social media platforms should not develop novel policies specifically for synthetic content.
• The most effective strategy is to strengthen the enforcement of current, general platform rules.
• Consistent application of existing policies, irrespective of content origin (human or machine), is recommended.