Anthropic has spent three years building its brand on a single, carefully tended premise: that it is the AI company that takes safety seriously. Its latest advertisement was supposed to prove it. Instead, it became the most-discussed AI marketing misfire of 2026 — and handed Sam Altman a gift he did not have to work for.
The ad, titled "There's Hope in Hard Questions", is the newest instalment of Anthropic's Keep Thinking campaign, created by agency Mother London and directed by Myles McAuliffe. It opens on footage of a burning house, then moves through facial recognition surveillance of crowds, a homeless person sleeping on the street, and a row of tombstones that multiple viewers quickly identified as Arlington National Cemetery. A voiceover poses questions: "Can AI be trusted?" and "Who's gonna hit the brakes if we need to?"
The intended message was that Anthropic grapples honestly with AI's dangers, and that grappling itself is a form of trustworthiness. The reaction suggested most viewers saw something else entirely.
How the Backlash Unfolded
The ad aired during the World Cup quarterfinal between Argentina and Switzerland over the weekend of July 12–13. Within hours, the cemetery imagery had gone viral in the worst possible way. One widely-shared post read: "When we raise the question of stopping a dangerously powerful superintelligence, we show 300 American gravestones for half a second." Another commenter summed up the paradox bluntly: "I can't stress enough how fucked up it is that Anthropic is running an ad that includes this image asking 'Who's gonna hit the brakes if we need to?'"
OpenAI CEO Sam Altman weighed in with characteristic brevity: "i thought this was satire, kept looking for the handle to be spelled c1audeai or something." For a company whose entire brand proposition rests on being the serious adult in the room, being publicly mistaken for a parody account by your chief competitor is a uniquely damaging outcome.
Critics noted a structural problem with the ad's logic. The spot transitions, roughly at its midpoint, from doom imagery to a more hopeful register — questions like "Could AI help people stop feeling misunderstood?" and "Can AI help me be a better teacher?" The pivot was meant to land as reassurance. For many viewers, it landed as whiplash. One Instagram commenter described it as the ad "attempting to shift to optimism in the second half to dupe people in its slick messaging."
A Playbook That Worked — Until Now
The failure is particularly notable given where Anthropic's marketing was just five months ago. In February 2026, the company aired a 60-second pregame ad and a 30-second in-game spot during Super Bowl 60 built around a single sharp premise: OpenAI had announced it would start testing ads within ChatGPT, and Anthropic was making a public promise that Claude would remain ad-free. The tagline — "Ads are coming to AI. But not to Claude." — was irreverent, specific, and genuinely competitive. It won the Film Grand Prix at Cannes Lions 2026, the festival's top prize.
Those Super Bowl spots also earned the smoldering resentment of Altman, who called them "deceptive" — which, in the context of AI industry marketing, essentially functions as a five-star review. The contrast with "There's Hope in Hard Questions" could hardly be sharper. The Super Bowl campaign identified a concrete rival behaviour (ads in a chatbot) and skewered it. The new ad identifies a vast, ambient sense of civilizational dread about AI and... asks questions about it. From the company that builds the AI.
Fast Company put the Catch-22 succinctly: make an AI ad too fun and it trivialises serious issues; make it too self-serious and you're accused of being disingenuous. "There's Hope in Hard Questions" went for self-serious and overshot badly.
The Credibility Problem Underneath
The deeper issue is that the ad's conceit — Anthropic as the company willing to ask hard questions — runs directly into its recent track record. In February 2026, Anthropic quietly dropped a key safety pledge that had previously committed the company to halting training on an AI system if it could not guarantee adequate guardrails. And despite Dario Amodei's carefully cultivated reputation as AI's most safety-conscious CEO, reporting emerged that Claude was being used to select strike targets in military operations in Iran — a fact that sits uncomfortably alongside cemetery imagery and voiceovers about accountability.
Futurism put the tension plainly: "By being the ones that doomsay about AI the most, Anthropic can also present itself as the only company that can be trusted to develop it. Its actual track record suggests otherwise." The ad did not invent this critique, but it handed critics a very vivid illustration on which to hang it.
There is a version of this campaign that works. The Keep Thinking platform — launched in September 2025 via the original Mother London hero film — does contain a genuinely differentiated brand idea. Where competitors promise speed and productivity, Anthropic positions AI as a tool that should slow you down enough to think. The original film earned Anthropic a Best B2B Campaign award from Ad Age's 2026 Creativity Awards, and the company's revenue trajectory through that period — from $1 billion in annualised revenue in late 2024 to $30 billion by April 2026 — demonstrates that the positioning had real commercial resonance.
But "There's Hope in Hard Questions" extends the platform in a direction that strips away its intelligence. The original film reframed AI's risks as invitations rather than threats. The new ad just… shows the threats, then asks who's responsible for them, then pivots to slightly less bleak questions. It is doomerism as a brand exercise, and viewers — including many who work in tech and are already sceptical of AI safety posturing — were not inclined to be charitable.
What the Tonal Whiplash Reveals
The broader pattern is worth naming. Anthropic is a company navigating multiple simultaneous tensions: it must attract enterprise customers while warning about existential risk; it must compete commercially with companies it publicly critiques; it must market a product whose harms it has spent years foregrounding in policy documents and public statements.
When that tension is acknowledged wittily — as in the Super Bowl spots — it becomes a brand asset. When it is played straight, as in the cemetery ad, it becomes a credibility problem. The company that asked "Who's gonna hit the brakes?" is the same company that has not, itself, hit the brakes.
Altman's "satire" comment stung because it was accurate in a specific way: the ad looked like the kind of thing a parody account would produce to illustrate the gap between what safety-focused AI companies say and what they do. Anthropic didn't need a parody account. It produced the illustration itself.
The ad's tagline — "There's hope in hard questions" — is not wrong as a sentiment. The execution just made it very easy to ask the hardest question of all: if Anthropic is the one asking, should we trust the answer?
