Today’s digest looks at Anthropic’s decision to sue the US Defence Department, Meta being told by its Oversight Board that deepfake moderation is falling short and failing to handle the spread of misinformation, as well as a new addition to the Barb Ads Hub.  

Anthropic sues US Defence Department

Anthropic has filed two lawsuits against the US Department of Defence (DoD), following it labelling the AI giant as a “supply chain risk” – the first designation of its kind against a US company. 

This designation came as a result of a feud between both parties regarding Anthropic’s attempts to implement safeguards against the US military potentially harnessing the company’s AI models for mass domestic surveillance purposes or autonomous weaponry. The DoD also laid out a demand that any company doing business with the government cut all ties with Anthropic. Axios reported that the White House is currently preparing an executive order which formally instructs the federal government to remove it from all of its operations.

Anthropic alleges that its designation as a “supply chain risk” is unlawful, that it violates its rights under the First Amendment. Its California lawsuit reads: “These actions are unprecedented and unlawful. The constitution does not allow the government to wield its enormous power to punish a company for its protected speech”. 

However, Anthropic states its wish to stay committed to providing AI for national security purposes, and to continue its negotiations with the government. An Anthropic spokesperson told the Guardian, “Seeking judicial review does not change our longstanding commitment to harnessing AI to protect our national security”.

Meta’s deepfake moderation falling short  

Elsewhere in the realm of AI, Meta is under scrutiny for failing to provide robust enough methods for the identification of deepfakes on its platforms. The Meta Oversight Board, a semi-independent body tasked with guiding the company’s content moderation protocols, has called on the company to overhaul how it surfaces and labels AI-generated content. 

The Board concluded that “Meta’s current system to properly label AI content is overly dependent on self-disclosure of AI usage and escalated review and does not meet the realities of today’s online environment”. It also acknowledged the challenges that emerge from the cross-platform proliferation of this type of content, where the content may have originated on another platform such as TikTok. 

Meta is being pushed to improve its existing rules on misinformation, to establish a separate community standard for AI-generated content, and to develop better AI detection tools, among other measures.  

Barb adds pre-campaign functionality 

Barb, the UK’s joint industry committee and standard for TV measurement, has announced the addition of a pre-campaign functionality to its recently launched Barb Ads Hub (previously known as the Advanced Campaign Hub). The move brings campaign forecasting capabilities to the analytics system, allowing advertisers to forecast the unduplicated reach and frequency achieved by their campaigns on both linear TV and VOD services across multiple screens. 

The new hub was launched in response to clients seeking end-to-end campaign reporting capabilities to understand the media-measurement loop. 

According to Luca Vannini, Head of Campaign Audiences at Barb, these developments mean that the hub “is now the principal destination for close to 600 UK buyers and sellers of advertising for full-cycle campaign optimisation, with all data compiled and reported in line with joint-industry standards.”