Global Exclusive | Meta internal memo: 'Regulatory debate around AI will intensify in coming months'

Meta expects that the regulatory debate around artificial intelligence (AI) safeguards will intensify in the coming months, a top executive of the social networking giant wrote in a recent internal post. This was after two United States senators shot a letter to Meta chief executive officer Mark Zuckerberg in early June, raising concerns about an alleged leak of details pertaining to one of its AI models.

The memo was posted by Meta chief product officer Chris Cox and talked about the Menlo Park, California-based company’s plans for the second half of 2023. A copy of it has been reviewed by Moneycontrol. We have shared queries with Meta on the internal memo and we will update the copy when they revert.

In February, Meta had opened access to the code of LLaMA to vetted AI researchers, but it was soon found to have become downloadable by anyone on the internet on platforms such as BitTorrent. The Facebook parent has said that it would continue with the release strategy of LLaMA code to approved researchers as it allows the company to balance ‘responsibility’ and ‘openness’.

Subsequently, US senators Richard Blumenthal and Josh Hawley, who sit on the Subcommittee on Privacy, Technology, & Subcommittee on Privacy, Technology & the Law, asked the tech giant what steps it had taken since the release of LLaMA to prevent or mitigate damage caused by the dissemination of the AI model.

Hot button issue

In his internal post last month, Cox referred to the safeguarding of young users on Meta’s platforms as a hot-button issue, following an order from the US antitrust watchdog in May that proposed barring the company from monetising the data of minors.

“We still have our work cut out for us: we expect key regulatory challenges to advance meaningfully around the world and we will need to continue to engage in discussions about how best to safeguard teens on our platforms globally as youth becomes an increasingly hot-button issue,” said the note by Cox.

ALSO READ | Meta's Quest for Business targets 80,000 paid subscribers by year-end

In addition to barring Meta from making money on minors’ data, the United States Federal Trade Commission (FTC) has also alleged the company failed to comply with a 2020 privacy order. The competition watchdog said Meta "had repeatedly violated its privacy promises" and its "recklessness has put young users at risk”.

At the time, Meta fought back saying that the FTC diktat was "a political stunt" and that the watchdog had failed to act against "Chinese companies, like TikTok."

The sentiment against monetisation of minors’ data has also been echoed in India’s draft digital personal data protection bill that prevents tech companies from tracking or behavioural monitoring of children and tells them not to target advertisements at them.

Moneycontrol reported earlier that industry bodies representing tech companies like Meta, Google, Twitter, Apple and Microsoft have sought a revision in the definition of a child under the Digital Personal Data Protection (DPDP) Bill to mean an individual under the age of 13. The current definition of a child is anyone under the age of 18.

Encryption test

The internal post by Cox also said that Meta will continue on the path toward end-to-end encryption (E2EE) across its networks, delivering E2EE for the Facebook network and starting public tests on Instagram.

“As of now, a user has to opt into making a specific chat with another user on Instagram end-to-end encrypted. Meta is trying to move towards a future where all chats on Instagram are end-to-end encrypted, similar to WhatsApp,” said a person in the know.

In 2021, Meta-owned WhatsApp had sued the Indian government to block new legislation that mandated over-the-top platforms like itself to share information regarding the first originator of messages, a demand that the company claimed would require it to break its E2EE model.

Last year, the country’s government drafted another set of rules under a new telecom bill that could require internet platforms to intercept or disclose any message to public authorities.

“E2EE on Instagram is likely to place the company in regulatory crosshairs in some major markets like India. My sense is that the company will either fight it out on principles either on all fronts or no fronts at all,” the source added.