X Touts Grok for Brand Safety
In a leaked internal deck, Elon Musk's X is touting its 'Grok' AI model in a push to reassure advertisers about brand safety. The move follows backlash over inappropriate content on the platform, but highlights the ongoing risks for brands relying on any single platform's black-box AI for content moderation.
The push to position Grok as a brand safety solution follows a significant advertiser exodus from X, which began after Elon Musk's 2022 acquisition. Major brands like Apple, Disney, Coca-Cola, and IBM paused or stopped advertising due to concerns over their ads appearing next to pro-Nazi and other extremist content. This advertiser pullout was projected to cost X as much as $75 million in revenue by the end of 2023 alone. The brand safety issues were amplified by Musk's own endorsement of an antisemitic conspiracy theory in November 2023, which directly led to companies like Disney, Warner Bros. Discovery, and Paramount Global suspending their ad campaigns. A report from Media Matters, which found ads from major brands next to pro-Nazi content, further fueled the advertiser backlash. In response, X filed a defamation lawsuit against Media Matters, claiming the report was a malicious attempt to drive advertisers away. X's new pitch to advertisers claims that by using Grok to review posts and user profiles, it has achieved nearly 100% "brand safe" scores, as measured by third-party verification firms IAS and DoubleVerify. The system is designed to block ads from appearing alongside content from users who frequently post about sensitive topics and can target thousands of keywords and author handles. This move is part of a broader strategy to automate advertising on the platform, with Musk envisioning Grok eventually handling everything from ad creation to campaign optimization. However, the touting of Grok for brand safety comes shortly after the AI model itself generated controversy for creating "deepfake" sexualized and violent images, including of women and children, leading to regulatory scrutiny. This incident highlighted the ongoing challenges and risks of relying on AI for content moderation, particularly a model that has a "spicy" mode designed to produce edgy content. In a reversal of previous rhetoric, X's presentation to advertisers also promoted the use of "blocklists," a tool that allows brands to prevent their ads from appearing on specific accounts. This is a notable shift, as Musk had previously criticized and taken legal action against advertisers for using such tools. Despite the new safety push, advertiser trust remains low. A Kantar survey showed only 4% of marketers believe X provides a brand-safe environment, and a quarter of advertisers planned to cut spending on the platform in 2025. U.S. ad revenue for X was reported to be down 60% in September 2023, and the platform is expected to account for less than 1% of worldwide digital ad revenue.
