~The AI + Marketing Weekly~

what you need to know this week in ai + marketing in 15 seconds:

The AI visibility paradox: beloved ≠ findable

Four major publications this week (WSJ, Adweek, Rolling Stone, Search Engine Land) wrote: brands can be culturally iconic and still invisible to AI.

Here’s what’s happening:

Some brands are slapping “no AI” disclaimers on everything (Aerie led the charge). “Human-made” is becoming a trust signal - basically “no added sugar” for marketing.
But other beloved brands aren’t showing up in AI answers at all because they’re not structured for models to read, cite, or summarize.
Even when brands are “present” in AI, they’re often not recommendation-ready. The model hedges instead of choosing them because the positioning is vague or the problem they solve isn’t clear.

The bottom line: Being iconic in human terms isn’t enough anymore. You also have to be legible to machines. And if you can’t say what problem your brand solves in one clear sentence, AI won’t either.

This is old positioning advice wearing new clothes. But now the stakes are higher - because the machine won’t be charitable and fill in the gaps like a human might.

What else?

OpenAI acquired TBPN - fast-growing tech talk show and media business. AI companies buying distribution and narrative control, not just talent (OpenAI)
AI didn’t break marketing, it exposed what wasn’t working - weak positioning, bloated processes, shallow content strategies look worse when machines speed them up. AI is a stress test (Fast Company)
Don’t let AI destroy the skills that make your company competitive - automating wrong work can erode judgment, craft, institutional knowledge. Efficiency is great until you automate your own competence away (HBR)
Americans’ AI use increases while views on it sour - using AI more while feeling worse about it. 7 in 10 think AI will cut jobs. Adoption and affection are not the same thing (Quinnipiac)
125,000 users blocked Bluesky’s Attie app in first few days - new app lets you build custom social feeds by describing what you want. Second most blocked account on Bluesky behind JD Vance, lol (Bluesky)

🗞️ top ai + marketing news

brand trust, visibility, and the anti-slop reaction

Brands adopt “no AI” disclaimers to stand out amid the slop
The Wall Street Journal | ~4-min read
More brands are explicitly telling customers they are not using AI-generated imagery in ads, partly because consumer skepticism is growing and “human-made” is starting to work like a trust signal. Aerie is one of the clearest examples.
My take: “No AI” is starting to sound a lot like “no added sugar.” Marketing loves a label once it becomes a shortcut for quality.

Brands beloved by people risk being invisible to AI
Adweek | ~4-min read
The article’s point is blunt: a brand can be culturally loved and still barely show up in AI answers if it is not structured in a way models can read, cite, and summarize.
My take: Being iconic in human terms is no longer enough. Now you also have to be legible to the machines -ahhhh.

Why AI hedges about your brand instead of recommending it
Rolling Stone | ~4-min read
This piece argues that brands are often visible but not recommendation-ready in AI systems. Content alone is not enough; the model also needs confidence, clarity, and enough context to say what problem the brand actually solves.
My take: Plenty of brands are technically “present” in AI. Fewer are clear enough to be chosen (Similar sentiment as the above article.)

If you can’t say what problem your brand solves, AI won’t either
Search Engine Land | ~5-min read
A useful reminder that AI summaries tend to flatten vague brands. If your company cannot be described simply and specifically, the model will either hedge or default to generic category language.
My take: This is old positioning advice wearing new clothes. Still true.

The new rules of trust in an AI era
Fast Company | ~5-min read
As AI becomes more embedded in customer-facing experiences, trust depends more on clear disclosure, predictable behavior, and fewer surprises. People are not asking for less tech. They are asking for less weirdness.
My take: Most trust problems start with “wait, what is this doing?”

How to navigate brand authenticity in the age of AI slop
Entrepreneur | ~6-min read
This focuses on authenticity as an operational choice, not a slogan. Brands can use AI behind the scenes, but the more customer-facing the output becomes, the more careful they need to be about quality, tone, and honesty.
My take: Most audiences do not care that AI helped. They care when the result feels lazy.

search, agents, and machine-readable brands

Build your marketing ark: A framework for AI, empathy, and design
Search Engine Land | ~5-min read
This is really a framework for staying useful when AI changes how people discover and evaluate brands. The emphasis is on empathy, structured design, and creating experiences that work for both humans and AI systems.
My take: The title is a little grand, but the core idea is on it: brands need more than automation. They need coherence.

Agentic AI discovery requires machine-readable brands
MarTech | ~5-min read
As agents become more involved in search and shopping, brands need structured data, clear attributes, and consistent signals across the web. Otherwise they become harder for AI systems to recommend confidently.
My take: We are entering the era of “did the robot understand your PDP.”

What are AI agents? What marketers need to know
Sprout Social | ~6-min read
A practical explainer on what agents are, what they can do, and how they differ from regular generative AI tools. Useful for teams that keep hearing “agentic” and quietly hoping the meeting ends soon.
My take: Half the value here is just giving people a shared definition.

the stack, org design, and the work underneath the hype

The AI myths marketers believed and what the data actually shows
Exploding Topics | ~6-min read
Survey data pushes back on a few common assumptions, especially around maturity, ROI, and how many teams are truly beyond the experimentation phase.
My take: “Everyone is doing it” remains one of marketing’s least reliable sentences.

Don’t let AI destroy the skills that make your company competitive
Harvard Business Review | ~6-min read
HBR warns that automating the wrong work can quietly erode the judgment, craft, and institutional knowledge that made a company good in the first place. Efficiency is great right up until you automate your own competence away.
My take: Some work is worth keeping close, even if a spreadsheet says otherwise.

AI didn’t break marketing. It exposed what wasn’t working.
Fast Company | ~4-min read
This piece argues that AI is less a disruption than a stress test. Weak positioning, bloated processes, shallow content strategies, and bad measurement all look worse when the machine speeds them up.
My take: AI has a rude way of revealing where the work was already flimsy.

The promise and problem of AI in advertising
Marketing Brew | ~4-min read
A straightforward look at where AI is making advertising faster and cheaper, and where it creates new risks around sameness, governance, and quality control.
My take: The promise is always speed. The problem is usually taste.

From hierarchy to intelligence
Block | ~4-min read
Block is framing its org changes around a move away from hierarchy and toward faster decision-making with more intelligence built into the system. It reads like an operating model argument as much as a tech one.
My take: Every company wants to sound flatter and smarter.

AI mandates are a demand for cognitive surrender
Product Picnic | ~5-min read
A pushback on top-down AI mandates that treat tool use as a proxy for progress. The essay argues that forcing adoption without judgment, context, or room for disagreement can hollow out the work.
My take: “Use AI more” is not a strategy.

media, content, and who gets to shape the conversation

OpenAI acquires TBPN
OpenAI | ~3-min read
OpenAI acquired TBPN, the fast-growing tech talk show and media business, saying it wants to accelerate the global conversation around AI while preserving TBPN’s editorial independence. The deal also explicitly calls out comms and marketing value, not just audience reach.
My take: AI companies are starting to buy distribution and narrative control.

labor, security, and the wider AI mood

Hackers are posting the Claude Code leak with bonus malware
Wired | ~4-min read
Wired reports that leaked Claude Code materials are being reposted alongside malware, which is exactly the sort of ugly side-effect you get when hype, leaks, and opportunistic attackers meet in one place.
My take: Nothing says “hot product category” like a leak immediately turning into a malware distribution method.

Americans’ AI use increases while views on it sour
Quinnipiac University Poll | ~4-min read
Quinnipiac found that Americans are using AI more while feeling worse about it. Seven in 10 respondents said they think AI will cut jobs, and younger people were especially pessimistic.
My take: Adoption and affection are clearly not the same thing.

Oracle layoffs continue as AI spending rises
CNBC | ~4-min read
Oracle is making cuts while ramping AI infrastructure spending, part of a pattern where AI is both the justification for big investment and the backdrop for workforce reduction. I could not access the full CNBC piece directly, but recent reporting ties the layoffs to Oracle’s AI cloud build-out and broader cost pressure.
My take: The AI boom is generating two very different feelings at the same time depending on which side of the org chart you’re on.

Sycophantic AI decreases prosocial intentions and promotes dependence
Science | ~5-min read
New research points to a nasty social effect of sycophantic AI systems: they can reinforce bad behavior and make users more dependent, especially in advice-like contexts. I could not access the full paper directly through the browser, but the abstract and coverage point in the same direction.
My take: “The bot agrees with me” is maybe not the healthy product feature some people think it is.

👩‍💻 thought-leader highlights

Culture Slant – “5 things I’ve learned about AEO for brand building”
AEO isn’t just SEO rebranded. Brands that win are cited, not just ranked. Write clear, definitive answers, build authority signals, and optimize for how LLMs quote you - not just how Google ranks you. Thank you for sharing this with me!

Hamna Aslam Khan – “Best Claude Code setups from Reddit”
Power users are turning Claude into full workflows: project folders + memory + skills + connectors. The takeaway: performance comes from setup, not prompts.

Charlie Hills – “The complete Claude guide: Chat, Cowork, Code”
Clear breakdown: Chat = thinking, Cowork = execution, Code = systems. Marketers should move from writing prompts to building repeatable pipelines.

Animalz – “AEO glossary”
AEO is quickly becoming its own discipline. This glossary is a good reset on how AI search works—from citations to retrieval—so teams can align on what actually drives visibility in LLMs.

🛠️ latest ai + marketing tools

Netflix VOID
Netflix just open-sourced a tool that removes objects from video, and unlike every other tool, it also fixes what those objects were doing to the scene. Remove a person holding a guitar, and the guitar falls.

Canva Magic Layers
Turns any flat image into a fully editable, layered design inside Canva, so the AI visual you generated but couldn’t change is now something you can actually tweak without starting over.

Claude unified directory
You can now share skills you’ve built with teammates and your whole org in one click. Skills, connectors, and plugins all live in one place in the new directory. Available on Team and Enterprise plans - and signals Claude is building toward a full platform, not just a chat product.

Google Gemma 4
Google’s new open-source AI model family, free to use commercially, runs on your phone or laptop, no cloud required. Built on the same research as Gemini 3, just without the paywalls.

Slack gets 30 new AI features
Slackbot can now take meeting notes across any video platform, work outside of Slack on your desktop, and act as a lightweight CRM, its biggest update since Salesforce bought it for $27.7B. Powered by Claude.

Bluesky launches Attie (and it is not beloved…)
A new app that lets you build your own social feed by describing what you want to see - no coding, no settings. 125,000 users blocked it in the first few days, making it the second most blocked account on Bluesky behind JD Vance. (LOL)

Google Veo 3.1 Lite
Google’s most affordable AI video model yet, at less than half the cost of its previous tier with the same speed - and timed perfectly after OpenAI quietly shut down Sora.

OpenAI Codex goes pay-as-you-go
Teams can now add Codex-only seats to their workspace with no fixed monthly fee - you only pay for what you use, and OpenAI also dropped the standard Business seat price from $25 to $20.

Microsoft Copilot Researcher gets smarter
Copilot now uses two AI models on every research task - one writes, a second reviews and fact-checks before you see the final report. Uses both Claude and GPT under the hood. A “Council” mode shows you where the two models agree or disagree side by side.

Contralabs
An evaluation platform where real creative professionals judge AI model outputs head-to-head - measuring taste, not just accuracy. Useful benchmark if you’re trying to figure out which AI tools actually produce good creative work.

Unwrap
AI that automatically surfaces what your customers are actually saying across every channel, so you’re not manually digging through dashboards to find the patterns. Backed by Atlassian Ventures, used by Lyft, JetBlue, and Perplexity.

Beams by Teleport
A secure way to run AI agents inside your company’s systems - each agent gets its own permissions and leaves a full audit trail. MVP launches April 30. More dev-facing, but worth knowing if your team is building anything with agents.

💼 cool ai + marketing jobs

Google - Product Marketing Manager, Industry and Applied AI

EliseAI - Growth Strategy, Manager | Housing

Adobe - Campaign Content Manager, Applied AI

Scrunch - AI Strategist

ElevenLabs - B2B Marketing Lead - ElevenCreative

Snap Inc. - AI Community Program Manager

Thanks for reading! This post is public so feel free to share it.

Reply anytime and let me know:

Topics you’d like covered in future weeks
Sections you love or could do without
Questions about AI, marketing, or anything else

Love y’all,
Carley

Carley Lake

📺 Open AI bought a talk show