Who is paying creators to shape the AI debate?
What's the deal? In her September 5, 2026 video, Sabine HossenfelderDealroom has a profile for this one. Try Dealroom → says she was offered money to produce a video warning audiences that artificial intelligence could destroy humanity. She declined after the proposed brief included the video’s subject, sentences to read, and references to quote. Her broader argument is that both sides of the AI debate are paying creators to make a case: the sponsorship itself is not necessarily improper, but the audience should be able to see who funded the message and whether the sponsor influenced the editorial content.
Why now? Hossenfelder points to several examples from the video. She says the Center for AI Safety has handed out grants for creators to produce AI-safety content, while Control AI has sponsored videos presenting AI as an existential threat. She also describes the Future of Life Institute’s “Protect What’s Human” campaign, which invites people to follow campaign directions, submit drafts for feedback, publish social-media content, and receive payment. On the other side, she cites paid creator activity connected with OpenAI, Anthropic, Google, and Perplexity, as well as a reported $5,000 offer from Build American AI for a TikTok presenting China’s AI progress as a threat to US jobs.
What could go wrong? The risk is not limited to ordinary advertising. When a creator’s sponsorship is undisclosed or difficult to identify, a paid message can appear to be independent science or technology reporting. Hossenfelder says sponsors may seek creators who are already sympathetic to a particular position, creating influence without explicit control over the final script. She references reporting by Taylor Lorenz for Wired and a Washington Post report that the Future of Life Institute funded 30 projects to develop AI-safety content. These claims describe reported programmes and allegations; they are not, by themselves, proof that every creator or organisation mentioned acted improperly.
The signal: Nirit Weiss-Blatt, replying to Hossenfelder on X, adds that the spending is not symmetrical and argues that “the doomers are way ahead”. The useful takeaway is a media-literacy rule rather than a verdict on either camp: when content makes strong claims about AI’s benefits or dangers, check the sponsor, the disclosure, the campaign brief, and whether the creator distinguishes paid advocacy from independent analysis. Hossenfelder describes herself as an “info-encer” whose aim is to give viewers information and let them think for themselves.
Read more: Science with Sabine · Sabine Hossenfelder on X · Nirit Weiss-Blatt on X · Sabine Hossenfelder’s website
Key points
What's the deal? In her September 5, 2026 video, Sabine Hossenfelder says she was offered money to produce a video warning audiences that artificial intelligence could destroy humanity. She declined after the proposed brief included the video’s subject, sentences to read, and references to quote. Her broader argument is that both sides of the AI debate are paying creators to make a case: the sponsorship itself is not necessarily improper, but the audience should be able to see who funded the message and whether the sponsor influenced the editorial content.
Why now? Hossenfelder points to several examples from the video. She says the Center for AI Safety has handed out grants for creators to produce AI-safety content, while Control AI has sponsored videos presenting AI as an existential threat. She also describes the Future of Life Institute’s “Protect What’s Human” campaign, which invites people to follow campaign directions, submit drafts for feedback, publish social-media content, and receive payment. On the other side, she cites paid creator activity connected with OpenAI, Anthropic, Google, and Perplexity, as well as a reported $5,000 offer from Build American AI for a TikTok presenting China’s AI progress as a threat to US jobs.
What could go wrong? The risk is not limited to ordinary advertising. When a creator’s sponsorship is undisclosed or difficult to identify, a paid message can appear to be independent science or technology reporting. Hossenfelder says sponsors may seek creators who are already sympathetic to a particular position, creating influence without explicit control over the final script. She references reporting by Taylor Lorenz for Wired and a Washington Post report that the Future of Life Institute funded 30 projects to develop AI-safety content. These claims describe reported programmes and allegations; they are not, by themselves, proof that every creator or organisation mentioned acted improperly.
The signal: Nirit Weiss-Blatt, replying to Hossenfelder on X, adds that the spending is not symmetrical and argues that “the doomers are way ahead”. The useful takeaway is a media-literacy rule rather than a verdict on either camp: when content makes strong claims about AI’s benefits or dangers, check the sponsor, the disclosure, the campaign brief, and whether the creator distinguishes paid advocacy from independent analysis. Hossenfelder describes herself as an “info-encer” whose aim is to give viewers information and let them think for themselves.
Read more: Science with Sabine · Sabine Hossenfelder on X · Nirit Weiss-Blatt on X · Sabine Hossenfelder’s website
