A KitKat Heist & Big Tech on Trial

Welcome to The Uncertain Times. This week the stories pulled in three directions at once – and all of them, somehow, arrived at the same place.

AI is getting more personal, commercial and contested. A jury in Los Angeles handed big tech its biggest reckoning yet. Italy’s regulators went after beauty brands for marketing skincare to children. Wikipedia banned AI-written articles. And someone stole 12 tonnes of KitKat.

The thread running through all of it: accountability.

This week, we unpack:

The personalisation pivot: Google connects your Gmail to its AI
ChatGPT reaches $100m in ad revenue
Why AI-generated content is collapsing in Google rankings
The world’s largest encyclopedia bans AI-written articles
Meta and Google found liable for harm to a child
Cosmeticorexia: Italy investigates Sephora and Benefit over skincare marketing to children
The great chocolate heist: 12 tonnes of KitKat vanish – and the brand response was flawless

The personalisation pivot: Google just connected your Gmail to its AI

Search has always known what you typed into the search bar. Now it knows what you booked, what you bought and who you emailed last Tuesday.

Google has expanded Personal Intelligence across AI Mode, Gemini and Chrome in the US, moving it beyond beta into broader consumer use. The feature connects Gmail, Google Photos, Calendar and other first-party apps so that AI-generated responses draw on personal context – not just the prompt in front of it.

Ask for a travel itinerary and it’ll reference your hotel confirmation. Ask for a product recommendation and it’ll factor in what you bought last month.

That makes results harder to replicate, rank against or track, especially in AI Mode, where outputs may vary based on user history, purchases and behaviour. Two users asking the same question will increasingly get different answers, shaped by different histories.

The implications for brand visibility are significant. Being present in AI-generated results was already tricky to measure. Being present in personalised AI-generated results is a different challenge entirely.

The StrategiQ takeaway: As AI search becomes personal, brand authority (the kind built through consistent thought leadership, earned media and third-party corroboration) becomes more important. Be the brand that genuinely shows up in the right conversations over time.

ChatGPT hits $100m in ad revenue and self-serve opens in April

OpenAI’s nascent ads business surpassed $100 million in annual recurring revenue less than two months after launching its US pilot. That’s from less than 20% of eligible users seeing ads daily, meaning inventory is about to get significantly larger. Now its expanding into Canada, Australia and New Zealand.

Self-serve advertiser capabilities are scheduled for April, opening access to the broader advertising market beyond the 600+ brands currently in the pilot.

The significance here isn’t the revenue figure – it’s what it signals. ChatGPT is becoming a mainstream ad surface faster than most marketing strategies have accounted for.

The StrategiQ takeaway: The brands that figure out AI advertising will have a meaningful head start. That said, high-intent, conversational environments reward relevance and specificity. The brief needs to be different.

Quality over quantity: AI content is collapsing in Google rankings

Here’s some data that should land on every content strategy desk this week.

A study by SE Ranking, covered by Search Engine Land, tested what actually happens to AI-generated content in Google search over time. The short version: Google indexes it quickly, and then demotes it. Pages without authority, unique insight or trust signals saw rankings collapse within months. Only 3% of AI-generated pages remained in the top 100 after three to six months.

This is Google doing what it’s always done – rewarding content that proves its worth through engagement, authority and genuine usefulness. AI-generated filler passes the initial indexing test and fails the durability one.

The finding reinforces something we’ve been saying for some time: volume is not a strategy. The brands winning in search (and in AI-generated responses) are the ones investing in content that is differentiated, expert and genuinely worth reading.

The StrategiQ takeaway: If your content strategy has shifted toward volume over quality in the last 12 months, this data is your course correction. AI is a useful tool for speed and structure. It is not a substitute for original thinking, real expertise or a distinctive point of view. The 3% that holds its ranking has something to say. The 97% that doesn’t was just filling space.

Wikipedia says no: The world’s largest encyclopedia bans AI articles

Wikipedia has banned the use of AI-generated text by its editors, stating that the use of LLMs to generate or rewrite article content is prohibited. The vote passed 40 to two among the site’s editors. Limited exceptions remain for copyediting and translation, but the content itself must be human.

The reasoning goes beyond editorial standards. Wikipedia is one of the world’s most visited websites and a primary source of training data for AI models. LLM-generated content on Wikipedia presents a compounding risk: inaccurate or hallucinated text enters the encyclopedia, gets scraped by AI companies and re-enters future model training data.

In other words, AI writing Wikipedia articles would eventually mean AI learning from AI, a loop that degrades the quality of the very data that underpins the tools we’re all using.

This matters for brand strategy too. Wikipedia remains one of the most trusted sources that LLMs draw on when characterising companies, categories and reputations. A platform committed to human-verified accuracy is a healthier foundation for the AI ecosystem than one full of generated content.

The StrategiQ takeaway: Wikipedia’s decision is a useful reminder that the sources AI learns from shape the answers AI gives – and that the quality of those sources is everyone’s problem.

Big tech’s big tobacco moment: Meta and Google found liable for harm to a child

A Los Angeles jury found Meta and Google negligent for designing social media platforms that are harmful to young people, in a $6 million verdict that could be the tipping point for thousands of similar cases.

The case involves a 20-year-old woman who said she became addicted to YouTube at the age of six and Instagram at nine. Internal Meta documents presented in court showed executives describing efforts to attract and retain young users, including a memo stating: “If we wanna win big with teens, we must bring them in as tweens.”

This is Meta’s second significant loss in US courts this week, following a New Mexico jury finding the company guilty of concealing information about the risks of child sexual exploitation and the harmful effects of its platforms on children’s mental health.

Many commentators are calling it big tech’s big tobacco moment. That comparison is worth sitting with. The tobacco industry knew. It concealed. It was eventually held to account at enormous scale. The social media industry is now on a similar trajectory and the legal and regulatory consequences are only beginning.

The StrategiQ takeaway: For brands advertising on these platforms, this is a moment to review both your placement strategy and your brand association. The reputational and regulatory pressure on Meta and Google’s core platforms isn’t going away.

You may also like: Australia’s social exodus

Sephora, Benefit and the children’s skincare problem

There’s a word you need to know this week: cosmeticorexia.

Italy’s competition authority has launched investigations into LVMH-owned Sephora and Benefit Cosmetics over claims their marketing has fuelled an unhealthy obsession with skincare among young girls, including children under ten. The alleged unfair commercial practices link to the use of young micro-influencers to promote adult serums, face masks and anti-ageing creams to a pre-teen audience.

The ‘Sephora kids’ trend has been building for years -TikTok and Instagram full of children documenting elaborate skincare routines, hauling retinol serums and creams designed for adults. Sephora boasts nearly 23 million Instagram followers and over two million on TikTok, with the brand sitting at the centre of tween beauty culture.

The platform dynamics that made that possible – algorithmic amplification, influencer economics, the blurring of entertainment and advertising – didn’t create cosmeticorexia, but they accelerated it. And now a regulator has decided that acceleration looks like liability.

LVMH says it operates in compliance with applicable regulations and will cooperate fully. The more interesting question is whether compliance and responsibility have been treated as the same thing, and whether this investigation is the first of many across Europe.

The StrategiQ takeaway: Read alongside the Meta and Google verdict, this is part of the same story: regulators across multiple jurisdictions are moving toward holding platforms and brands accountable for the downstream effects of engineered engagement with young and vulnerable audiences. The legal risk is real. The reputational risk is greater. Brands that have relied on the grey area between targeting children and marketing to them are going to find that area shrinking fast.

Someone stole 12 tonnes of KitKat and the brand’s response was perfect

A giant chocolate bar one week, KitKats the next.

Nestlé recently confirmed that 12 tonnes of KitKat bars (413,793 units to be precise) were stolen in transit between a factory in central Italy and their destination in Poland. The truck and the merchandise remain unaccounted for…

The story is inherently enjoyable. Pre-Easter chocolate heist. Hundreds of thousands of bars potentially circulating across European markets. KitKat has asked anyone who finds a matching batch number to alert the company, and confirmed there are no consumer safety concerns and no risk to supply.

But what’s worth noting here isn’t the theft, it’s the response.

“We’ve always encouraged people to have a break with KitKat,” said Nestlé, “but it seems thieves have taken the message too literally and made a break with more than 12 tonnes of our chocolate. Whilst we appreciate the criminals’ exceptional taste, the fact remains that cargo theft is an escalating issue for businesses of all sizes.”

That’s a brand statement doing three things at once: owning the moment with humour, reinforcing a fifty-year-old brand asset without forcing it and landing a serious point about cargo crime without losing the tone. It doesn’t feel like it was written by a committee. It feels like someone had fun and trusted the brand enough to let them.

The StrategiQ takeaway: A crisis (even a mildly absurd one) is a brand moment. How you respond tells people more about who you are than most of your carefully planned campaigns. KitKat leaned into its distinctive asset under pressure and came out with a story that travelled globally within 48 hours. That’s not luck. That’s a brand with enough clarity about what it stands for that the right response was obvious.

The StrategiQ takeaway

This week’s stories are all, at their core, about accountability.

The common thread: the systems that shape what people discover, believe and buy are being contested – legally, editorially, commercially. The businesses navigating that well aren’t waiting for the dust to settle. They’re building brand authority that holds up regardless of how the landscape shifts.

The technology moves fast but the strategy still has to lead.

Because…

IT’S ALWAYS STRATEGY.

Sources:

Personalised AI, Big Tech on Trial & a 12-Tonne KitKat Heist