Containing disinformation

Initiatives by digital giants will be critical

The Supreme Court on Thursday stayed the notification regarding the establishment of a fact-checking unit under the Press Information Bureau till the Bombay High Court decides the petitions challenging the amendments to the Information Technology Rules. Meanwhile, digital giants Google and Meta are finetuning strategies for combating election disinformation. Apart from releasing new detection tools, and offering training to fact-checkers, both will collaborate with fact-checking initiatives run by other organisations, and coordinate with the Election Commission of India (ECI). While all the major digital platforms mounted individual efforts against disinformation with mixed results in the 2019 general elections, the 2024 campaign against disinformation is much more coordinated. Over 40 large multinational corporations signed an accord at the Munich Security Conference, agreeing to cooperate globally to fight election-related disinformation in 2024. The threat perception is very high, since artificial intelligence (AI) can now generate realistic audio-visual content featuring politicians. The digital giants will coordinate with sundry news organisations to try and limit misinformation, limit interference with voters, and enhance transparency and accountability on the platforms they control.

This is significant since Meta owns Facebook, WhatsApp, and Instagram, while Google owns YouTube and its dominant search engine. The strategies involve connecting with independent fact-checkers and vernacular content creators and publishers, offering collaborative platforms to share fact-checks, research resources, and alerts on misinformation and deepfakes, aiming to prevent such content from going viral. Meta already removes misinformation such as content designed to suppress voting, or content that threatens or encourages violence. It claims a network of independent fact-checkers covering 15 Indian languages. It has a WhatsApp helpline to report AI-generated misinformation, especially deepfakes, which helps with reporting and verifying suspicious media. Google will create a common repository for publishers to tackle misinformation at scale. Fact-checks in multiple languages and formats, including videos, will also be shared and amplified via Google’s partners. Both companies will also impart training in advanced fact-checking methodologies and deepfake detection, and roll out detection tools like Google Fact Check Explorer and Meta Content Library. Fact-checkers may label content as “Altered”, which will lead to it being filtered out. This could include faked, manipulated, or transformed audio, video, or photos.

Facebook already labels photorealistic AI images by putting visible markers and embedding invisible machine-readable metadata. It claims to be building tools to detect and label such AI-generated content from Google, OpenAI, Microsoft, Adobe, and Shutterstock (content from these sources is commonly posted on Facebook, Instagram, WhatsApp, and Threads). Meta also requires advertisers to disclose when they use AI to create content centred on political or social issues. It bans ads with content debunked by fact-checkers, or ads designed to discourage people from voting. If the campaign against disinformation is to be effective, it will be necessary to detect and remove fake content in real time to prevent its viral spread before it is flagged. It is also important to utilise such tools without bias; it can be easy to misuse fact-checking tools to selectively block content from a particular political faction. Given the stakes, it is to be hoped these measures (and similar ones undertaken by platforms such as X or other search engines) will be effective, and deployed in good faith, without fear or favour.