The Growing Threat of AI-Generated Disinformation in Nigeria

Nigeria is facing a rising challenge with AI-generated disinformation, particularly as it prepares for its next major election cycle in 2027. Recent incidents demonstrate the evolving sophistication of these tactics:

  • An image showing Peter Obi meeting deceased military dictator Sani Abacha circulated widely on social media despite being fabricated.
  • A TikTok video falsely claimed President Bola Tinubu was critically ill after being poisoned, shortly following the death of former President Muhammadu Buhari.
  • Deepfake clips have been used to manipulate public opinion by falsely attributing endorsements or policy positions to candidates.

Google’s SynthID Detector: A Promising Tool with Limitations

Google recently launched its SynthID Detector as a free platform to identify AI-generated content. The tool works by scanning for imperceptible watermarks embedded in images, videos, and audio files created using compatible AI systems. While this technology has potential, experts note several limitations:

  • It only detects content with Google’s watermark or those from partner platforms.
  • Content generated using open-source tools, mobile apps, or AI systems outside the partnership will bypass detection.
  • The tool currently supports only English, limiting its effectiveness in a multilingual country like Nigeria where political discourse extends to Hausa, Yoruba, Igbo, Pidgin, and other languages.

Existing Efforts to Combat Disinformation

Nigeria has taken steps to address this challenge:

  • INEC (Independent National Electoral Commission) established an Artificial Intelligence Division to manage AI-related risks during elections.
  • The Center for Collaborative Investigative Journalism developed ElectionWatch, an AI-powered tool tracking misinformation on TikTok and Telegram.
  • FactCheckAfrica’s MyAIFactChecker offers credibility assessments in multiple languages, though it currently lacks deepfake audio verification (with upgrades planned).

Why 2027 is Different

The ease of creating convincing deepfakes has dramatically increased since the 2023 elections. What once required specialized technical skills can now be done with free online platforms in minutes.

This evolving threat requires a comprehensive approach that includes technological solutions, media literacy initiatives, and collaboration across sectors to ensure Nigeria’s democratic processes remain secure.