Jul 29, 2025
BRICS Nations Push for Stronger Data Protections Against AI Misuse
At the recent BRICS summit in Rio de Janeiro, leaders from Brazil, Russia, India, China, and South Africa gathered to discuss the future of artificial intelligence (AI). They made it clear that AI should be developed carefully, with stronger data protections to protect data and prevent misuse, especially by tech giants that are located in wealthier nations.
At the recent BRICS summit in Rio de Janeiro, leaders from Brazil, Russia, India, China, and South Africa gathered to discuss the future of artificial intelligence (AI). They made it clear that AI should be developed carefully, with stronger data protections to protect data and prevent misuse, especially by tech giants that are located in wealthier nations.
A major focus of the discussions was on how personal and private data are used without consent to train AI models. BRICS leaders expressed concern about the way many companies scrape content from the internet to power AI tools. A draft statement seen by Reuters suggested that the bloc is calling for a mechanism to make sure content creators are paid fairly. It also aims to limit the amount of data collected without justification.
Indian Prime Minister Narendra Modi spoke about the need for global rules to check the accuracy of digital content. He also shared that India will host an “AI Impact Summit” in 2026 to continue these discussions.
Modi’s message aligned with the overall goal of the summit: AI should support human values, not harm them. He stressed that developing countries should also have a say in how AI is managed and controlled.
Key BRICS AI Goals
Countries must ensure consent before regional data is used for AI.
Unique contributors like languages or cultures should be compensated.
Develop open-source multilingual, multicultural AI tools.
Invest in shared ethical and safety AI frameworks
The bloc will also work with the UN AI Advisory Body to help shape global AI governance policies. This united approach from BRICS could influence future talks on how to manage AI globally, including meetings at the UN and G20.
📰 Mini Headlines
📺 Nigeria Fines Africa’s Largest Pay TV Operator for Data Breach
MultiChoice Nigeria, the country’s biggest pay TV provider, has been fined ₦766 million (about $500,000) by the Nigeria Data Protection Commission for illegally transferring user data across borders and violating privacy rights. The breach affected not only subscribers but also individuals indirectly linked to them. Regulators found the company’s data practices “patently intrusive” and disproportionate.
FlirtAI App Leaks 160K Private Chat Screenshots
FlirtAI, an iOS dating assistant app, exposed over 160,000 private chat and profile screenshots through an unsecured cloud storage bucket. Many of the leaked conversations involved teens, which has raised serious concerns about consent and privacy. The app’s design encourages users to upload screenshots for AI-generated replies. However, it failed to protect sensitive data. The developers have since secured the bucket but haven’t issued a public statement.
Germany Moves to Ban DeepSeek Over Privacy Violations
Germany’s data protection authority has asked Apple and Google to remove the Chinese AI app DeepSeek from their stores, citing illegal data transfers to China. Officials say DeepSeek failed to prove that user data is protected under EU standards. The app reportedly collects extensive personal information, and its privacy policy lacks transparency. Other EU countries, including Italy, have already taken similar action.
HCL Software Launches Sovereign AI for Government Data Privacy
HCLSoftware has introduced Domino IQ. It is a sovereign AI platform that aims to help governments and regulated sectors protect sensitive data. The launch is part of Domino 14.5 and supports compliance with laws like the European AI Act. It allows organisations to use trusted AI models while avoiding foreign cloud dependencies. Over 200 government agencies are expected to benefit from enhanced privacy and security features.
