What we're saying. What's being said.
We publish what we learn — and we read widely.
What we’re reading
Fortune · Oct 13, 2026
Strict liability is gaining traction as an idea for forcing AI companies to get serious about safety
Fortune: Jeremy Kahn on strict liability — holding AI companies responsible for harm their models cause regardless of negligence — gaining support as a lever for safety, with no federal AI regulation in place (nuclear-plant operators and aircraft makers already face it in some cases). The NYT editorial board backed it; a Forecasting Research Institute survey of 157 experts, superforecasters, AI-risk specialists, and members of the public found US-only strict liability would cut catastrophe risk more than a federal licensing regime — at least in half. The downsides: liability only comes after harm, a catastrophic judgment could bankrupt a company (so who pays victims?), open-source models are hard to pin down (many vendors in China), and it could push AI firms to withhold models and compete with their own customers. Kahn’s conclusion: make it one layer of a broader approach — standards, insurance, licensing, inspections — not the whole answer.
What we’ve written
Patrick Adams · Oct 15, 2026
IKEA’s CMO on AI: it distributes your brand, it doesn’t create it — and only a consistent message survives the handoff
On LinkedIn: Patrick on IKEA’s global CMO telling Digiday that AI should 'distribute your brand, not create your brand.' His read: IKEA can use AI well because it has said the same thing for years — beautiful, affordable furniture for the many — while brands that keep shifting their message to chase short-term sales will struggle, because the models can’t work out what they stand for. AI assistants build their answer from your website, press, support, and financial filings; inconsistent messaging across them sends mixed signals. The prep is internal: agree on what the company stands for and say it the same way everywhere. His test — write the one sentence you want an AI assistant to say about your company, then check whether your website, your support team, and your latest press release all match it. Most brands would fail.