Microsoft and OpenAI pool $2 million to combat deepfakes in elections

A $2 million (Rs 14.7 crore) fund has been set up by Microsoft and OpenAI to fight election deepfakes. The project aims to combat the threat that artificial intelligence (AI)-generated disinformation poses during the next elections, which are predicted to involve two billion voters in 50 nations. Notably, political parties have been advised not to use deepfakes in digital campaigns by the Election Commission of India.

The project aims to protect communities that are more likely to accept manipulated content without questioning its authenticity. The tech companies are strategizing a unified approach specifically aimed at tackling deepfakes designed to mislead voters. Other major AI firms have begun addressing these risks by implementing restrictions in their software. For example, Google's Gemini AI chatbot will not answer election-related queries, and Meta is also limiting election-related responses via its AI chatbot.

OpenAI recently introduced a deepfake detector tool for disinformation researchers. This tool is designed to identify fake content created by its own DALL-E image generator. In addition, OpenAI has joined the steering committee of the industry group known as Coalition for Content Provenance and Authenticity (C2PA). The C2PA includes members like Adobe, Microsoft, Google, and Intel.

The newly established "societal resilience fund" is part of a broader movement toward responsible AI.

Microsoft and OpenAI aim to "further AI education and literacy among voters and vulnerable communities". The fund will offer grants to many organizations, including Older Adults Technology Services (OATS), Partnership on AI (PAI), C2PA, and the International Institute for Democracy and Electoral Assistance (International IDEA).