In May, TikTok announced that automatic labeling of AI-generated content on its platform. However, this does not apply to all versions of the application. new report from the Mozilla Foundation and AI Forensics found that TikTok’s Lite-Save Data version, which is aimed at users in poorer markets, not only does not flag AI-generated content, but also does not employ other, similar safeguards.
“Labeling is a very important tactic that platforms use to provide some form of trust and security,” says Odanga Madung, a Mozilla contributor and co-author of the report.
For example, users of the full version of TikTok will see labels indicating that the content is graphic or depicts unsafe behavior. Some content on topics like elections and health, similarly, includes a notification encouraging users to access trustworthy information through an in-app “resource center.”
TikTok Lite has none of these guardrails. Among other things, this means that, in the face of dishonest AI-generated content that’s been a problem in elections around the world, users in poorer markets are given less information about what’s imitation and what’s real than users in richer markets.
Madung wonders why, of all the features that could be cut to optimize the app, the company included those that make the platform safer for users. “We don’t know if that’s a choice or just negligence,” he says.
“This report contains several factual inaccuracies that fundamentally misrepresent our approach to safety,” a TikTok spokesperson said in a statement. “The fact is that content that violates our policies is removed from TikTok Lite in the same way as it is from our main app, and we offer numerous safety features.” The company declined to point out any specific inaccuracies.
Lite versions of apps have long been a way for companies to gain market share in areas where users face high data costs or can only afford less advanced phones. In 2015, Meta, then Facebook, launched Facebook Lite, a stripped-down version of its app that was more compatible with 2G data networks. That same year, it also launched Free Basics, which allowed users in the Global South to access the platform and some other sites without paying data charges. (An app or service that meets these criteria is called a “zero-day.”) At the time, the project was widely criticized, particularly in India, for creating a second-class experience for poorer customers.
TIK Tok shot in 2018 in Thailand as a Lite app and quickly expanded to other markets in Southeast Asia, including Indonesia, Malaysia and the Philippines. The app, which unlike the full TikTok app, can run on 2G and 3G networks, now has over 1 billion downloadsaccording to data from the Google Play store. (TikTok Lite is only available for Android phones.)
“Most users in the Global South are low-income and resource-poor,” says Payal Arora, professor of AI inclusive cultures at Utrecht University. The simplified versions of the apps aid companies bring these people on board, which she says is even more essential than in the past because “data is the currency in this AI-driven, AI-hungry market.”