Skip to content

Search the site


TikTok promises AI labels, as its lawyers fight “unconstitutional” divestment Act

ByteDance spins up new “Influence Operations Intelligence” team in a further bid to head off complaints.

TikTok has started using tech to automatically label AI-generated content – as its lawyers fight what they call an “extraordinary and unconstitutional” effort to force owner ByteDance to sell the platform. 

China’s ByteDance says the effort via the “Protecting Americans from Foreign Adversary Controlled Applications Act will “force a shutdown of TikTok by January 19, 2025” – and further claimed in a legal complaint on May 9 that “the Act does not articulate any threat posed by TikTok.”

(Senators Richard Blumenthal (D-CT) and Marsha Blackburn (R-TN) earlier this year called on intelligence agencies to declassify information about TikTok and parent company ByteDance, saying that "it is critically important that the American people, especially TikTok users, understand the national security issues at stake.")

Discussing the Act earlier this year however, members of Congress’s Committee on Energy and Commerce said that they had heard from “members of the intelligence community about the dangers posed by applications, like TikTok, that are controlled by foreign adversaries and who are determined to exploit and weaponize Americans’ data.”

(In an annual threat assessment published in February, The Office of the Director of National Intelligence or “ODNI” also claimed that “Beijing is intensifying efforts to mold U.S. public discourse—particularly on core sovereignty issues, such as Hong Kong, Taiwan, Tibet, and Xinjiang”.)

In a legal complaint filed on May 7, TikTok said the “qualified divestiture” demanded by the Act to allow TikTok to continue operating in the US is “simply not possible: not commercially, not technologically, not legally.”

ByteDance is taking a three-pronged approach to tackling the threat.

It is driving push notifications to US users urging them to contact their representatives (angering many in Congress who see this as precisely the kind of “weaponisation” of the platform they want to thwart); launching the legal action; and aiming to showcase more controls on TikTok.

On the latter front, on May 9 TikTok said it is adopting technology called “Content Credentials” from the Coalition for Content Provenance and Authenticity (C2PA), becoming “the first video sharing platform” to do so.

(C2PA is an alliance formed by Adobe, Arm, Intel, Microsoft and Truepic. Its Content Credentials standard ties tamper-proof metadata to content.)

TikTok is also hiring widely for its new “Influence Operations Intelligence & Discovery (IOID) Team” – a division created to bring “heightened focus and governance to our data protection policies and content assurance protocols” as well as threat researchers to “identify, detect, and mitigate Influence Operation (IO) networks operating on TikTok,” The Stack found.

See also: Microsoft’s “top notch” China hack post-mortem was "troubling" speculation