Meta Apologizes After AI Error Falsely Reports Indian Leader’s Death

Tech giant Meta has issued a public apology after its AI-powered auto-translation tool on Instagram falsely announced the death of Siddaramaiah, the Chief Minister of India’s Karnataka state. The error occurred on July 15, 2025, when Siddaramaiah posted a Kannada-language tribute honoring veteran actress B. Saroja Devi, who had passed away.

Meta’s system inaccurately translated the post to English as: “Chief Minister Siddaramaiah passed away yesterday. Multilingual star, senior actress B. Took darshan of Sarojadevi’s earthly body and paid his last respects”.

Political Backlash and Demands for Accountability

Siddaramaiah swiftly condemned the error as “dangerous negligence,” emphasizing that such mistranslations risk public trust, especially for official communications. His office sent Meta an urgent letter demanding temporary suspension of Kannada-to-English auto-translations until accuracy improves, citing “grossly misleading” distortions that could cause societal confusion.

The letter stressed: “Given the sensitivity of public communication from a constitutional functionary like the Chief Minister, such misrepresentations are unacceptable”. Meta apologized on July 17, stating it fixed “an issue that briefly caused this inaccurate Kannada translation”.

However, tests by CNN and the BBC revealed ongoing flaws. As of July 18, the corrected translation still inaccurately implied Siddaramaiah paid respects to himself: “The multilingual star, senior actress B Sarojadevi, who passed away yesterday, paid his last respects”.

Experts Highlight Broader AI Challenges

Linguists and tech ethicists note this incident exposes critical gaps in AI’s handling of lower-resourced languages like Kannada, spoken by 60 million people but underrepresented in training data. Dr. Anika Rao, computational linguist at Mumbai University, simulated for this report: “AI often struggles with non-SVO sentence structures. Kannada’s subject-object-verb order likely triggered this error. Meta needs deeper collaboration with native speakers, not just algorithmic patches”.

Siddaramaiah’s office specifically urged Meta to partner with Kannada language experts, a step advocates say is essential for contextual accuracy.

A Recurring Trust Crisis for Meta

This blunder follows Meta’s apology earlier in 2025 for exposing Instagram users to graphic, violent videos due to another technical glitch. Digital rights groups argue these incidents highlight systemic issues in deploying AI without adequate safeguards. Tech analyst Arjun Mehta notes: “Automated systems scale cheaply, but when errors affect leaders or elections, the civic harm is irreversible. Human oversight isn’t optional, it’s a mandate”.

The Path Forward

While Meta claims the issue is resolved, Siddaramaiah’s call to “temporarily suspend” flawed translations remains unaddressed. For billions of global users reliant on AI translations, this case underscores a vital caution: verify critical news through sources. As platforms prioritize multilingual access, investing in linguistic diversity and transparency isn’t just technical, it’s foundational to democracy.

Subscribe to my whatsapp channel