Microsoft Pledges AI Safeguards to Help Protect Nigeria’s 2027 Elections from Deepfakes

Microsoft has pledged to incorporate the Independent National Electoral Commission’s (INEC) final list of candidates for the 2027 general elections into its systems to help prevent the generation of misleading AI-created images of political candidates.

The commitment emerged from the Artificial Intelligence and Democracy Forum (AIDF) 2026, held in Abuja from October 14 to 16 under the theme “Will Algorithms Determine the 2027 Vote? Safeguarding Nigeria’s 2027 General Election in the Age of Artificial Intelligence.” The forum was convened by Yiaga Africa, INEC, the Centre for Journalism Innovation and Development (CJID), and Dataphyte Foundation, with support from organisations including Microsoft, Google, TikTok, Luminate and the MacArthur Foundation.

According to the forum’s outcome document, more than 350 participants, including representatives of electoral bodies from Senegal, Liberia, Sierra Leone, The Gambia, Guinea and Togo, took part in discussions on protecting electoral integrity against AI-enabled disinformation.

INEC outlined several measures, including finalising an AI governance framework for elections, adopting a binding Human Oversight Principle, deploying AI-powered cybersecurity protections against denial-of-service and brute-force attacks, and establishing an Electoral Information Integrity Coalition. Under the proposed oversight principle, critical electoral decisions involving voter accreditation, result collation and candidate qualification would not be left entirely to automated or autonomous systems.

INEC Chairman Professor Joash Amupitan, SAN, declared that “Algorithms will not, and must never, decide the 2027 vote,” warning that manipulating voters through false information could pose a serious threat to the election involving more than 103 million registered voters and 176,846 polling units. The forum’s outcome document also made 13 recommendations, urging political parties to adopt an Ethical AI Code of Conduct, media organisations to establish digital forensic desks, and security agencies and the judiciary to strengthen their capacity to investigate cybercrime and verify digital evidence.

Technology platforms were also urged to label AI-generated political content, restrict deceptive candidate likenesses, remove synthetic impersonation promptly and improve content moderation in Nigerian languages. The recommendations further called for transparent and auditable AI-powered result-aggregation systems, while reaffirming INEC’s sole legal authority to declare final election results and warning that the spread of synthetic media could lead people to dismiss authentic evidence as fake.

At the forum’s innovation showcase, Election GPT, developed by Pawavox, won the AI in Elections Sandbox Judges’ Choice award in the Election Observation and Data category, receiving a ₦13 million prize and technical support to develop and scale the platform before the elections. The tool enables observers to submit geotagged, timestamped and cryptographically signed voice and video reports while flagging potentially manipulated content for human review; another platform showcased was the WaFlo Incident Board, developed by Crown Softech Ltd, an official Meta Tech Provider.