Age verification of teens tricky, needs flexibility: Meta VP

She was responding to this paper’s question about hardcoding the threshold for considering an individual a kid like India’s draft data protection Bill has, which has proposed to treat an individual below the age of 18 as a child. Social media platforms like Facebook and Instagram have been casted additional obligations to protect the data of these individuals under the draft Bill.

Instead, Davis said that there should be a “flexible” way around the issue where data collection can be minimised. “For example, facial recognition technology can be used that will allow us to identify age. I think it’s trying to find that implementation flexibility while understanding the need for standards that is the current conversation happening between policymakers and industry.”

Asked how tech companies are expected to react to growing demands from across the world on allowing access to encrypted services for violations like people sharing child sexual abuse material (CSAM), Davis said, “you can’t break encryption for one issue and not break it for another. You break it for one issue, you, you break it for everything”.

In 2021, when the Centre had notified the Information Technology Rules, 2021, it had said that services like WhatsApp will have to help law enforcement authorities on finding the “first originator” of a message, a requirement colloquially referred to as ‘traceability’. WhatsApp and Facebook have both challenged the provision in the Delhi High Court, saying that it undermines Indians’ right to privacy and could result in “a new form of mass surveillance”.

Across hearings in the case so far, the government has contended that WhatsApp needs to develop a technological measure to enable traceability, and it is necessary to create a safe cyberspace for citizens.

Davis said that just because messaging on some platforms is encrypted does not mean that there aren’t other ways to reduce or prevent the spread of child sexual abuse material on these platforms.

“WhatsApp takes down hundreds of thousands of accounts that it believes are sharing CSAM. They do it through a variety of different signals that don’t require that they break encryption – whether that’s using photo DNA across group photos that are signals to people because otherwise, how are they going to know what’s in that message themselves. They use the names of a particular group chat, so that they can identify that that is a chat where that (CSAM) may be occurring. So there are many, many ways that they can can identify potential child abuse accounts,” Davis said.

“We put a lot of money into training our system on a number of different languages. Our content moderators try to catch such accounts in 20 different Indian languages. We can always do better and do more, and we are committed to that. So we’re always evolving, whether it’s our technology, whether it’s adding additional languages, to try to do better. The approach has to be multilayered,” she added.