AI-driven ads will be hard to spot

Metacognition

In an article for The Conversation titled, “OpenAI looks to online advertising deal. AI-driven ads will be hard for consumers to spot,” Stuart Leeds, a professor of economics at the University of Leeds, took a thoughtful look at OpenAI and how online advertising could eventually manifest itself in the company’s answer engine, ChatGPT.

Mr. Leeds saw OpenAI’s deal with Shopify and AI’s unique “metacognition” as a catalyst.

He explained:

“On any given topic, ChatGPT will ‘know’ more about it than an average person. A particularly knowledgeable person might not get caught out. But nobody is an expert on everything, while ChatGPT can at least pretend to be (like any good salesperson).

AI large language models (known as LLMs) are also up to speed on the latest research on rhetoric, marketing and psychology. They can even identify deceptive sales techniques.

AI can also be tweaked to be persuasive. For instance, research has found that people are more likely to buy something when a salesperson or advert mirrors their personality. One study found that ChatGPT can accurately predict a person’s personality from relatively little information. Over time then, ChatGPT could be programmed to make predictions about us, and then start acting like us.”

So, could a 300×600 display ad from ChatGPT predict something the user not only wants, but really really really wants?

Read more on TechXplore. (September 4)

From tipsheet: In theory, once this “AI and ads” nut is cracked for answer engines, advertising is going to be so much better for the user and the advertiser versus the old “search engine” days. That said, Meta, Google and Amazon are already well on their way using AI for ads – it’s just not happening in an answer engine format, yet.


LLMs & CHATBOTS

Developments

  • China’s DeepSeek Preps AI Agent for End-2025 to Rival OpenAI (September 4) – Bloomberg (subscription)
  • Inside the Anthropic ‘red team’ tasked with breaking its AI models—and burnishing the company’s reputation for safety (September 4) – Fortune
  • Expanding economic opportunity with AI (September 4) – Fidji Simo on OpenAI blog
  • Microsoft’s Copilot Outpaces ChatGPT in US Mobile Growth, Data Shows (September 3) – Adweek

BRANDS

Ad tech AI glossary update

Fresh on the heels of the IAB’s AI use case map on Wednesday, Ad Age published, “Ad tech glossary—from ‘ad exchange’ to ‘vibe targeting.’”

Ad Age’s Garett Sloane described the glossary’s intent on LinkedIn:

“The ad tech lexicon needs an update now that ‘vibe targeting’ entered the chat. So, I compiled these terms that either raise eyebrows, controversy, or simple conversation. (Plus, Ad Age made a nice dictionary-inspired graphic to accompany it.)

Thanks to…

Scope3 and Brian O’Kelley for confirming ‘vibe targeting’ is indeed a thing.

Ari Paparo for calling SSPs like he sees them—’ad exchanges.’

And some other terms that are just popping up more frequently: agentic, composable, dynamic …”

See the glossary on Ad Age. (September 4 – subscription)


AGENCIES

WPP CEO Cindy Rose on AI

She’s only four days in to her new role, but after speaking at WPP global town hall “live from New York,” Cindy Rose, CEO of agency holding company WPP, addressed the wider universe and posted a letter on LinkedIn.

Ms. Rose was explicit that AI was one of three “core principles” of WPP going forward:

Harnessing our Al advantage: We’ve made significant investments in building market-leading Al capabilities and solutions, and we need to lead with these in every client conversation. We all need to level up our skills and become Al superusers and embrace these tools to power our daily work, to deliver efficiency and quality to our clients and delight them with cutting-edge innovation, creativity and business outcomes.”

Read the letter on LinkedIn. (September 4)


SELL-SIDE

Publishers adding answer engines

Some publishers are implementing their own answer engines with the goal of generating more revenue—prompting Adweek reporter Mark Sternberg to wonder from the user’s perspective:

“…why would anyone willingly restrict the scope of their search to a website, when an internet-wide search would necessarily include that website, plus a trillion others?

Since then, I’ve had some pretty compelling conversations.”

Taboola CEO Adam Singolda provided one of the compelling conversations and discussed his company’s DeeperDive answer engine product that is restricted to the publisher’s content and data only.

Mr. Singolda made the case for one version of the publisher-hosted, answer engine opportunity:

“All the traffic lost to answer engines is lost to answer engines, and tools like DeeperDive are not remedies for that. Instead, what they seek to do is mimic a new behavior that users are becoming increasingly accustomed to—the chatbot interfaces—and enabling that behavior on publisher websites. This way, whatever percentage of direct traffic a publisher has, the gizmo deepens their engagement, encouraging them to spend far longer on the page and offering up a gold mine of first-party data in the process.”

The other version is subscription and dependent on proprietary data or archives.

Read more in Adweek. (September 3)


TECH

More on AI ad network ZeroClick

AdExchanger talked to Ryan Hudson about the recent launch of his “reasoning-time ad network for AI,” ZeroClick.

Mr. Hudson previously co-founded Honey – an affiliate marketing startup acquired by PayPal for $4 billion in 2020.

Hudson told AdExchanger that ZeroClick’s origin is the Pie ad blocker which he described as a contextual ad targeting solution at its core. With ZeroClick, the Pie ad blocker technology can be used to read the context of AI tool output (like an answer engine or an AI note taker) on a publisher’s website and insert Zero Click-controlled ads.

AdExchanger’s James Hercher explained further:

“For instance, if a retailer uses ChatGPT as the underlying technology for its on-site or in-app chatbot, the chatbot’s results are not informed by any paid marketing resource. With ZeroClick, businesses would have more control over how their products or brands are presented and over the reviews referenced. They might also be able to pay to have their products show up more often.”

Read more on AdExchanger. (September 3)

More:

  • Why AI ad networks are popping up, even on ChatGPT (on ZeroClick’s launch from August 27) – Ad Age
  • Also, hear more about Honey and ZeroClick on last Friday’s Marketecture podcast here.
  • Honey has been involved in an affiliate links “overriding” controversy since last year. You can read more about it here on Wikipedia.

TECH

AI in marketing survey

A new survey from Basis Technologies of 140 respondents across agencies, brands and publishers says that there is good news and bad news when it comes to AI adoption.

Let’s start with “the good:”

“95% of marketing professionals use generative or agentic AI tools at least once a month – one-third are using it every day.

The most popular tasks for using AI according to 76% or more of respondents is ‘ideation/brainstorming’ and ‘research.’

The most-used AI tool for advertising work is ChatGPT, which is utilized by 88% of respondents.

No other platforms garnered more than 45% of respondents. The popularity is spread across multiple tools when asked about use of paid platforms.

More than 65% say their organizations currently pay for premium AI tools – a significant increase from 44% in 2024.”

“The bad” list begins with environmental concerns.

Read more in the release. (September 4)


TECH

AI search reducing QPS

Inuvo ad operations executive Eric Tilbury commented on U of Digital’s recent newsletter observations about two key shifts in advertising: “1) ad spend moving from display to CTV, and 2) AI search reducing the QPS that DSPs rely on for transactions.”

Tilbury continued on LinkedIn regarding what this means for the open web:

“The shift to CTV is driven by two factors. First, CTV is a strong channel worth investing in. Second, and less discussed, CTV allows continued ID targeting, keeping the ad tech spine intact.

Does this mean display ads are dead? No way. Are they less effective? Nope. But you need to adjust your approach. My data spine is built on content and consumption patterns, not user IDs, so it holds strong without processing the massive QPS that major DSPs need for buying signals. The future of open web display targeting will have far fewer stable IDs but will focus on specific URLs showing strong user intent…”

Read more from Mr. Tilbury on LinkedIn. (September 4)


PROMPT

Will digital advertising exchanges still be necessary in five years?

Response from Meta.ai:

Digital advertising exchanges will likely continue to play a crucial role in the industry, driven by advancements in AI, cross-channel integration, and privacy enhancements.

Here’s what we can expect:

  • Advancements in AI: Artificial intelligence will improve targeting precision, optimize bids, and detect fraud, making ad exchanges more efficient.
  • Cross-Channel Integration…

Read more on tipsheet.


MORE

  • Uber offers driver partners a side hustle: AI data labelling (September 4) – The Economic Times
  • Ex-Salesforce co-CEO Bret Taylor’s Sierra is the latest $10 billion —AI agents for customer service— startup (September 4) – CNBC
  • Warner Bros. Discovery sues AI giant Midjourney for copyright infringement In major legal battle (September 4) – The Hollywood Reporter
  • Eskimi pioneers AI-powered contextual advertising with DeepContext (September 4) – press release
  • DeepBrain AI launches automated UGC Ad Videos – press release