What it means for us
A weekly intelligence brief for leadership communicators, speechwriters and executive advisors.
What It Means For Us is my weekly AI intelligence brief specifically tailored for leadership communicators. It’s not about catching all the stories important to us – it’s about finding out what AI news is useful to us now and where it may be headed. I also occasionally review various AI tools. My views here are not purchased—they’re for your consideration and judgment. Content covers the previous week, contains specific notes for those restricted to internal AI models such as Copilot, and is delivered each Monday. Sign up here.
In a nutshell
This week, AI is less about what machines can do — and more about who has to clean up after them.
The EU’s marketing transparency rules land in August (sounds boring but directly impacts our work). Ford rehired the engineers it let go because, it turns out, a dashboard doesn’t know why your engine makes that “brumpf-huga-chugga-booma” sound (nor does your humble narrator). And a member of Congress (en los estados unidos) had to explain why the phrase “Claude responded” showed up in a legislative summary that was just spellchecked, honest.
If you’ve ever wondered “how do mistakes like this happen? Who’s responsible for reading this stuff over?” — well, you already know the answer: it’s our job as communicators now, whether we asked for it or not. Increasingly, we’re the last line of defense, asking “would the audience feel duped by this?” before it goes out the door.
Grab a coffee — there’s a lot to get through.
First, some (mostly) good news.
A founder uses AI to make cancer treatment more understandable and navigable
TechCrunch profiled Connor Christou, who used AI tools to organize and make sense of his own cancer journey — blood results, scan data, wearable outputs, personal notes, all of it. A few editorial points here. First, the idea is not (AI construction alert!) that AI is replacing doctors; it’s that some of the footwork can be done before we show up at the doctor’s office so that real, human doctors can make real, informed human decisions. Second, would it not be wonderful to live in a world where everyone had the resources (tech, cash, a decent national healthcare system) to do this?
The rest.
Thomson Reuters warns of an AI “value gap” in professional services
Thomson Reuters’ 2026 Future of Professionals report says AI is now embedded in professional work, but expectations and reality are drifting apart. 74% of professionals use AI several times a week, but more than a third are using unsanctioned tools their organization doesn’t know about. Sounds like a problem for the IT nerds to think about? Think again: this is about confidentiality, quality control and reputation. If your organization is using enterprise models (such as Copilot) and those tools are weak, people are going to try and get around them (clutch pearls at will). If employees are bringing their sensitive drafts, client materials and strategy language home and running it on ye-olde-free model, you’re going to be mopping up ye-olde comms disaster. Can we say it enough? Comms must be represented in the rooms where executive decisions are made. Full stop.
But that’s if your company is actually using AI and not just “into it.”
European Central Bank: AI adoption is broad, but serious use is still rare
ECB researchers found more than 70% of euro-zone firms report using AI — but only 7% use it intensely. The deepest users tend to be smaller, younger, service-oriented, and more likely to build custom tools than just buy licenses. Reuters’ summary is sober: adoption is everywhere, transformation is not. Keep this in mind when someone in your org is waving the AI flag because they figured out how to build a simple prompt. “We use AI” means almost nothing now. The better question is where AI is actually built into decisions, workflows, and outcomes; how much your org is spending on unused tools; and what value you’re truly deriving from them.
“Claude responded” appears where it really should not (or, “What — me worry?”)
From across the creek in America (what? I live in Geneva!), Rep. Anna Paulina Luna said her staff used Claude only for spelling and grammar on a defense-bill amendment summary, after screenshots circulated showing a stray “Claude responded” fragment left in the text. She denied AI drafted the actual legislative language and said formal bill text comes from House Legislative Counsel, which is barred from using AI. Maybe so — but nothing undercuts “careful public process” quite like leaving chatbot residue in a government document for the internet to find.
Ford’s AI quality lesson: institutional memory is not a dataset
Ford acknowledged it had to rehire, promote, or newly hire 350 veteran engineers — many of them former employees — after relying too heavily on AI-driven design and quality systems that weren’t trained on the judgment those engineers carried out the door with them. The payoff: Ford just took the top spot among mainstream brands in JD Power’s Initial Quality Survey for the first time in 16 years. The lesson isn’t that AI is useless. It’s that AI performs better paired with people who know the work deeply enough to judge what it produces. Somewhere, a veteran engineer looked at a dashboard reading and said the equivalent of “I told you that noise wasn’t normal.” And as a guy who used to work on the assembly line (GM, many moons ago), I can assure you none of this has shaken the core belief of those putting the vehicles together: that engineers and AI are both idiots. Look, the lesson here is the same as last week’s issue: before we automate the work, can we not first understand the work?
OpenAI, Anthropic, Microsoft and Amazon back a $500 million AI workforce initiative
Contractor alert: former U.S. Commerce Secretary Gina Raimondo and former Indiana Governor Eric Holcomb launched Raise Us, a $500 million initiative to help workers adapt to AI-driven workplace change. Backers include the OpenAI Foundation, Anthropic, Microsoft, Amazon, and IBM, plus state governments and employers. The program will pilot AI career coaching and retraining incentives across several states. Which reminds me — if your org is looking for an AI for Leadership Communicators Workshop, you know where to find me. Otherwise, here’s the story:
AI agents are becoming a new audience — and brands are scrambling to catch up
At a Cannes event, The Atlantic’s Alice McKown and Elf Beauty’s Ekta Chopra told Axios that brands are being told to restructure their content for AI agents that now shop, search, compare, and act on a customer’s behalf. McKown’s warning: publishers need to rethink licensing and access before AI companies intercept the value of their content. Chopra’s message was on point: these tools need conversational context, not bloody keyword soup — the average ChatGPT search has grown from five words to twenty-three. Key thing to keep in mind for communicators: the audience isn’t only human anymore. It’s an automated valet with purchasing power and zero patience for your PDF brochure. On a brighter note, can we retire PDF brochures? And while we’re at it, white papers and forwards? Asking for a friend.
“If an AI agent had to explain our organization, choose our product, or compare us with a competitor, what approved content would it actually find?”
EU AI Act transparency rules are about to land on marketing and comms desks
Woo hoo! The exciting topic of EU AI Act transparency rules! But don’t fall asleep yet. This stuff has consequences for communicators. Starting August 2, EU AI Act transparency obligations hit marketing and comms teams using generative AI. Article 50 is about whether people know they’re talking to AI or consuming AI-generated content — chatbot disclosure on first contact, provenance metadata, watermarking, content labels. Perhaps a bit more troubling: a lot of comms teams are already using AI in ways they haven’t formally classified. If your organization operates in or markets to the EU, this applies to you.
“Map one campaign from draft to publication and mark every point where AI touched the work. Then decide what needs disclosure, metadata, human review, or legal sign-off.”
One we missed — and worth knowing if you missed it too
New York’s first-in-the-nation law requiring disclosure when ads use AI-generated “synthetic performers” actually took effect, and was announced as in effect, back on June 9 — outside my usual weekly window, so it slipped past me at the time. It’s worth flagging now because it’s exactly the kind of thing that becomes a legal checklist while everyone’s still treating it as a LinkedIn debate. Advertisers operating in or reaching New York audiences must conspicuously disclose any AI-generated human likeness in a commercial ad. Civil penalties run $1,000 for a first violation, $5,000 after that. If your creative review process doesn’t already ask “does this contain a synthetic performer that needs disclosure,” now’s the time to add it.
Audit your organization’s “AI workforce story.”
Imagine you’re an employee who just heard the organization is expanding its use of AI. Based only on public statements, internal communications, and executive speeches, answer:
Then: identify the biggest gaps, and draft three messages that would actually reduce employee uncertainty.
AI for Leadership Communications · Geneva