cd /news/artificial-intelligence/the-trouble-with-ai-mania-show-me-th… · home topics artificial-intelligence article
[ARTICLE · art-119749] src=adexchanger.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

The Trouble With AI Mania; Show Me The Receipts

A new Adweek report finds that about 14% of ads appearing in ChatGPT have no relevance to the user's query, highlighting limited advertiser controls over ad placement in the AI chatbot. Separately, Omnicom transferred more than 460 engineering and product staff who built its Omni AI platform to third-party contractor Endava, raising concerns about outsourcing core AI development. Amazon confirmed it is removing specific product details from order confirmation emails to bolster privacy, a move that also directs customers to its app and website, while the Trump administration's Justice Department filed a statement of interest supporting OpenAI's fair use defense in the New York Times copyright lawsuit.

read4 min views1 publishedSep 3, 2026
The Trouble With AI Mania; Show Me The Receipts
Image: Adexchanger (auto-discovered)

Losing Control

Breaking news: AI is flawed!

As AI critics are apt to point out, the technology sometimes generates new problems, rather than just solving old ones.

Case in point: Ad buyers are showing interest in ChatGPT ads – but the channel is so new that advertisers have “limited controls” over what content their ads show up next to, Adweek reports. And, apparently, about 14% of ads in ChatGPT have no relevance to the user’s query. Although AI can be great for automating mundane tasks or for speeding up insights, there is such a thing as being too trigger-happy with AI. Not only can AI overreliance lead to less effective ads, it can also convince companies that they need fewer employees dedicated to customer service.

Omnicom, for instance, recently offloaded more than 460 members of the engineering and product teams that built Omni, its in-house AI platform, to a third-party contractor called Endava.

Omnicom says it has “entered a multi-year partnership with Endava” to support Omni’s development and operations. But, reportedly, the employees transferred to Endava will also be working on other projects beyond Omni support.

As one of the team members who was offloaded to Endava points out, “Is it really a good idea to outsource the whole engineering effort of what you’re trying to sell to clients?”

That Could’ve Been An Email

Amazon Prime customers are peeved over a new update the company made to its email systems.

The ecommerce behemoth is now stripping product details from order confirmation emails. Customers say they are seeing more generic descriptions in their receipts, such as “automotive item” in place of a specific car part.

Amazon tells Business Insider the purpose of the change is to bolster privacy protection by limiting the detail of any customer information that leaves Amazon’s four walls.

Amazon made a similar change to its emails in 2020 but eventually rolled it back. So why bring back order obfuscation now? It’s possible that Amazon doesn’t want AI tools to have easy access to its sales data.

Plus, if customers have to revisit Amazon’s site or app to check on their orders, they might end up spotting something else they want to buy. Indeed, Amazon acknowledges that the email change helps “direct customers to our app and website.”

It wouldn’t be the first time Big Tech claims privacy as the motive behind self-serving decisions. And it won’t be the last.

Who Saw This Coming?

Quelle surprise! The Trump administration is taking OpenAI’s side in the New York Times’ ongoing copyright lawsuit against the AI company, The Verge reports.

In a statement of interest filed this week, the Justice Department argues that OpenAI’s use of NYT materials to train its AI models is protected under fair use, mirroring OpenAI’s own argument.

That the Trump administration would side against the New York Times isn’t all that surprising, considering President Trump himself has personally criticized the newspaper plenty of times. He is also currently suing the NYT for defamation.

Not to mention, OpenAI has lots of ties to current government leadership. There’s the $200 million defense contract OpenAI secured in 2025, the $25 million in donations that Co-Founder and President Greg Brockman gave to Trump’s SuperPAC last September and the recent news that OpenAI might consider giving the government a 5% stake in the company, just for starters.

It’s also not the first time the DOJ has sided with an AI company in recent weeks. In August, US attorneys also argued in defense of Elon Musk’s xAI against a Minnesota ban on AI that lets users “nudify” images of people without consent.

However, when it comes to the NYT’s case against OpenAI, the final decision will be up to Judge Sidney Stein of the Southern District of New York – who was nominated by Bill Clinton in 1995.

But Wait! There’s More!

Amazon is adding a scam-detection tool to its AI shopping assistant. [TechCrunch] The branded content side of the creator economy might be in a slowly bursting bubble. [Digiday]

Uber plans to cut 10% of employees – though AI is apparently not a factor. [The Information] You’re Hired!

Media agency Crossmedia hires Ali Plonchak as chief growth and corporate strategy officer and promotes Bobby Singh to chief financial and operating officer. [release]

Here’s today’s AdExchanger.com news round-up… Want it by email? Sign up here .

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @adweek 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/the-trouble-with-ai-…] indexed:0 read:4min 2026-09-03 ·