Meta acknowledges gaps in algorithms during talks with Indian Govt

Meta admitted to problems with its algorithms during a technical meeting with India’s Ministry of Electronics and Information Technology (MeitY). Sources say the discussions focused on how Meta’s systems decide what users see, how content is moderated, and whether these measures protect users and comply with local laws.

This admission comes as the Government increases discussions with Meta about regulation and compliance. Talks began after a Facebook post by Prime Minister Narendra Modi was briefly restricted, but have since grown into a wider review of Meta’s main operations in India. The issue started with a video post by the Prime Minister on July 23, which was blocked for a short time due to concerns about exam paper leaks. Meta said the block was an “inadvertent error” by its AI filters, but MeitY was not satisfied with this explanation and called Meta’s global leaders for further talks.

Joel Kaplan, Meta’s Chief Global Affairs Officer, led a team to meet with IT Minister Ashwini Vaishnaw and Secretary S Krishnan.

Kaplan apologised on behalf of the company, called the removal an operational mistake, and promised stronger internal checks. Reports say Meta CEO Mark Zuckerberg also expressed regret about these issues.

The meetings lasted several days and included a technical session on how Meta’s algorithms and systems manage content delivery, recommendations, and moderation.

During the talks, Meta admitted there are gaps in its algorithms. Officials looked at how content is delivered, how moderation works, and whether recommendations are fair. The Government also asked about child sexual abuse material (CSAM or CSEAM), the spread of deepfakes and fake content, protections for verified and well-known accounts, and possible algorithm bias.

Meta executives reportedly admitted failures in handling CSAM and deepfakes, and said that spending a lot of money to promote certain types of content sometimes caused problems.

One main issue is Meta’s claim to protection as an intermediary under Section 79 of India’s Information Technology Act. These “safe harbour” rules usually protect platforms from being held responsible for user content if they act as neutral hosts and follow certain rules.

Indian officials argue that when algorithms actively select, recommend, and promote content, especially paid material, the platform is not just a passive host.

This challenges Meta’s usual defence and could mean the company will be held more responsible for harmful content.

The Government has made it clear that platforms in India must fully follow Indian laws and the constitution, no matter what standards they use in other countries.

Meta was told to list all problem areas and provide a detailed plan to fix them. Officials say they want ongoing improvements, not just quick fixes. Sources say Meta’s global technical team will likely stay in India for more talks.

More meetings, possibly led by the IT Secretary, are expected soon as both sides work to solve the remaining issues.

Both parties have agreed to keep working together until all concerns are addressed.

MeitY has repeated the need for quick action. For Meta, this shows that India’s regulatory stance is firm: market access depends on clear local compliance, strong safety systems, and transparency about how platforms shape public discussion.

This is not an isolated case. It is part of a global trend where governments want more control over the technical systems of social media giants. Whether Meta’s admissions and promised improvements lead to real changes in content safety, fairness, and accountability will be closely watched in the coming weeks.

The technical meetings continue, and the algorithms behind one of the world’s largest digital platforms remain under close review.