Domain-Specific Solution Proposal for AI-Driven Disinformation

A comprehensive study published in the Harvard Kennedy School Misinformation Review revealed that there is no single solution in combating AI-driven disinformation, and each sector must take layered measures suited to its own dynamics.

Scope of the Study and Expert Insights

In the study conducted by academics from Frankfurt University of Applied Sciences and IMT Atlantique, the opinions of 54 internationally serving experts were consulted to examine AI disinformation.

The research results published in July 2026 clearly demonstrated that the risks faced by different sectors diverge significantly in nature.

Threat Elements Varying by Sector

While video deepfake technologies are seen as the greatest danger in the political sphere, AI-generated written texts and fabricated studies pose the highest risk in the healthcare sector.

In the financial sector, voice cloning techniques rank among the riskiest methods, while 78% of experts defined election interventions via fake videos as the most urgent threat.

Experts Diverge on Precaution Measures

The study evaluated five different measures: government regulation, digital watermarking, media literacy, platform moderation, and technical detection tools.

Government regulation and media literacy ranked among both the most effective and least effective options, indicating that there is a debate on prioritization rather than polarization among experts.

Layered Approach Essential for the New Era

It was determined that public awareness tools currently reach audiences already proficient in technology and miss the segments truly at risk.

The researchers recommended strengthening source verification standards and platform moderation in the political sphere, labeling and medical infrastructure in healthcare, and authentication protocols in the financial sector.