The Challenges of Detecting Deepfakes in the Global South

When it comes to detecting deepfakes in the Global South, one of the biggest challenges lies in the quality of the media being used. Many detection models have been trained on high-quality media, which is not always the case in regions like Africa where cheap Chinese smartphone brands dominate the market. These phones often produce lower quality photos and videos, which can confuse detection models. Additionally, background noise in audio or compressing videos for social media can lead to false positives or negatives, making it difficult for models to accurately detect deepfakes in real-world scenarios.

Aside from generative AI, another common form of manipulated media in the Global South is cheapfakes. These are created by adding misleading labels or simply editing audio and video to mislead viewers. However, faulty models or untrained researchers may mistakenly flag these cheapfakes as AI-generated content. This can have serious consequences on a policy level, potentially leading legislators to crack down on imaginary problems. Inflating numbers of AI-generated content can be risky and could result in unnecessary restrictions.

Creating and running detection models for deepfakes requires access to energy and data centers, which are often lacking in many parts of the world. Without the necessary resources, it becomes nearly impossible to develop local solutions for detecting deepfakes. Researchers in regions like Ghana are left with limited options, such as using costly off-the-shelf tools, inaccurate free tools, or relying on academic institutions for access. The lack of local alternatives can lead to delays in verification processes, as sending data to other institutions results in significant lag time.

Impact on Information Ecosystem

The focus on detecting deepfakes may divert funding and support away from organizations that contribute to a more resilient information ecosystem. Instead of solely investing in detection technologies, funding should also be directed towards news outlets and civil society organizations that help build public trust. By supporting these institutions, it is possible to create a more robust environment for combating misinformation and disinformation.

The challenges of detecting deepfakes in the Global South are complex and multifaceted. From the quality of media to the access to tools and resources, there are numerous obstacles that researchers and journalists face in identifying manipulated content. By recognizing these challenges and investing in solutions that address the specific needs of these regions, it is possible to make progress in the fight against deepfakes and safeguard the integrity of information in the Global South.