According to a report from PPC Land this week, OpenAI is giving ChatGPT advertisers a new way to prove their campaigns work, and the clock is ticking to get ready.

As reported by PPC Land, the company announced that conversion-optimized campaigns will begin rolling out in early June, with early access opening June 5 for advertisers who have conversion tracking in place. The shift moves ChatGPT ads from a reach-and-traffic play into a direct performance channel, one where the platform optimizes delivery toward actual purchases and leads rather than just clicks.

The change is structural. Until now, ChatGPT’s Ads Manager offered two objectives: Reach and Clicks. Conversion optimization adds a third option, allowing the system to use post-click behavioral data to shift spending toward users statistically more likely to complete a defined action. That signal has to come from somewhere, which brings us to the infrastructure.

The measurement loop

OpenAI provides two tools to close the attribution loop: a JavaScript Pixel for browser-side tracking and a server-side Conversions API for direct backend-to-platform transmission.

The JavaScript Pixel captures a “privacy-preserving identifier” from the landing page URL and stores it in a first-party cookie. The Conversions API sends events server-to-server, which means browser-level ad blockers and cookie restrictions don’t interrupt the data flow. Running both simultaneously—a dual-tracking approach that mirrors how Meta’s Conversions API is commonly deployed—requires deduplication to avoid counting the same conversion twice.

For teams without dedicated engineering resources, a community GTM template published by marketing analytics professional Utku Gulden on May 7 offers the fastest path to pixel deployment through standard tag management workflows.

The performance question

Before June 5, comparing ChatGPT to other ad platforms meant looking at CPM rates and cost-per-click benchmarks. The conversion objective changes the comparison set. The relevant question becomes cost per conversion, which is the metric performance buyers use when deciding where to move budgets.

Early data is suggestive but unverified. Criteo, one of several technology partners operating inside the ChatGPT ad pilot alongside Kargo, Adobe, Pacvue, and StackAdapt, reported on May 5 that AI-referred conversion rates approached twice those of traditional search in several retail categories: consumer electronics, lifestyle and wellbeing, and home and garden. Those figures come from Criteo’s own client base. Independent third-party verification has not been published.

The platform crossed $100 million in annualized revenue within six weeks of launching, according to Reuters. With 700 million weekly active users, the audience scale has never been the question. Conversion optimization is what makes that audience actionable in the terms performance marketing actually uses.

The legal framework arrived before the technical one. OpenAI updated its US privacy policy on April 30, disclosing for the first time in binding language that it receives purchase data from advertisers and their partners to measure ad effectiveness, shares personal data with marketing partners for third-party targeting, and uses personal data to promote its own products. Five days later, the technical infrastructure the measurement model depends on was available at scale.

June 5 is the date the platform stops being a branding experiment and starts being a performance channel. Whether the conversion rates justify the move is a question each performance team will have to answer with their own data.

The post OpenAI is turning ChatGPT into a performance ad channel appeared first on The Media Copilot.

OpenAI today released GPT-5.5, what it says is its “smartest and most intuitive to use model yet, and the next step toward a new way of getting work done on a computer.”

The company says the model understands what users are trying to do faster, can carry more of the workload itself, and excels at writing and debugging code, researching online, analyzing data, creating documents, operating software, and moving across tools until a task is finished.

The company published performance numbers from Terminal-Bench 2.0, which tests complex command-line workflows requiring planning, iteration, and tool coordination.

OpenAI said GPT-5.5 outperformed its predecessor on every major coding and agent benchmark the company tested, while using fewer tokens and running at the same speed as the older model. On one third-party coding index, it matched leading rivals at about half the cost.

Keeping a larger model that fast required rebuilding inference as a single system rather than a patchwork of tweaks, the company said. GPT-5.5 was designed, trained and served on NVIDIA’s latest hardware, and OpenAI credited its own Codex tool and GPT-5.5 itself with helping hit the efficiency targets.

Early testers told the company the model seems to grasp how a codebase fits together — why something is failing, where the fix belongs and what else the change will touch.

Dan Shipper, Founder and CEO of Every, called GPT-5.5 “the first coding model I’ve used that has serious conceptual clarity.” After launching an app, he spent days debugging a post-launch issue before bringing in one of his best engineers to rewrite part of the system. To test GPT-5.5, he effectively rewound the clock: could the model look at the broken state and produce the same kind of rewrite the engineer eventually decided on? GPT-5.4 could not. GPT-5.5 could.

Pietro Schirano, CEO of MagicPath, saw a similar step change when GPT-5.5 merged a branch with hundreds of frontend and refactor changes into a main branch that had also changed substantially — resolving the work in one shot in about 20 minutes.

One engineer at NVIDIA with early access went as far as to say: “Losing access to GPT-5.5 feels like I’ve had a limb amputated.”

OpenAI is already running the model internally at scale. More than 85% of the company uses Codex every week across functions including software engineering, finance, communications, marketing, data science, and product management. The finance team used GPT-5.5 in Codex to review 24,771 K-1 tax forms totaling 71,637 pages, accelerating the task by two weeks compared to the prior year.

The model also shows gains on scientific and technical research workflows. On GeneBench, a new eval focusing on multi-stage scientific data analysis in genetics and quantitative biology, GPT-5.5 outperforms GPT-5.4 on problems that often correspond to multi-day projects for scientific experts. On BixBench, a benchmark designed around real-world bioinformatics and data analysis, it achieved leading performance among models with published scores.

In a notable example, an internal version of GPT-5.5 with a custom harness helped discover a new proof about Ramsey numbers — one of the central objects in combinatorics — later verified in the Lean proof assistant. The result is a concrete example of GPT-5.5 contributing not just code or explanation, but a novel mathematical argument in a core research area.

OpenAI says GPT-5.5 was released with its strongest safeguards to date, including tighter controls around cybersecurity workflows and repeated misuse patterns. The model was evaluated across the company’s full safety and preparedness frameworks, with input from nearly 200 trusted early-access partners before launch.

GPT-5.5 is available today in ChatGPT and Codex for Plus, Pro, Business, and Enterprise users. GPT-5.5 Pro is rolling out to Pro, Business, and Enterprise tiers. API access, which requires different safeguards, is coming “very soon,” OpenAI said.

The post GPT-5.5 Is ‘Our Smartest Model Yet,’ Says Company With History of Saying That appeared first on The Media Copilot.

OpenAI is taking the “New Year, new you” approach when it comes to its business strategy. To start 2026, the company said it would soon be introducing ads ChatGPT, which is a bit of a shock considering CEO Sam Altman had previously called advertising a “last resort” for the business.

Whether this is truly the end-of-the-road option is tough to judge without a peek at OpenAI’s balance sheet, but it’s not hard to see why they’re feeling pressure. After Google released Gemini 3 in the fall—which led to strong leaderboard results, increased market share, and plenty of AI-world praise—Altman reportedly declared a “code red” to keep ChatGPT best-in-class. And while OpenAI’s fundraising has been impressive, Google is a $4 trillion company. OpenAI needs all the resources it can get.

So ChatGPT users are getting ads. That’s a gamble, because plenty of signals suggest people don’t want advertising blended into AI answers. A report from Attest, a consumer research company, found that 41% of consumers trust AI search results more than paid search results, suggesting that AI users like that they don’t have to worry about ads in AI summaries, even if their accuracy may sometimes be questionable. Hallucinating is apparently less of an offense than selling out.

Still, the broader direction is hard to ignore: ads inside AI products are starting to feel less like a possibility and more like gravity. People don’t love commercials on TV or streaming, either, but they’re baked into the media economy. Google is already placing ads in AI Overviews and AI Mode, and it may eventually do something similar in Gemini—even as executives insist there are no plans. Whatever happens with the Gemini chatbot specifically, Google looks committed to threading advertising through its AI experiences, which is about as surprising as rain.

The ad pie gets sliced thinner

For publishers, none of this sounds like a party. If OpenAI becomes an ad platform, it’s another Big Tech competitor fighting for the same finite pool of digital ad dollars—right alongside Google, Meta, and Amazon. Meanwhile, the traffic equation keeps getting uglier: chatbots summarize the web, and that often removes the reason to click. There’s a reason web traffic to publishers dropped by a third last year.

But there’s a twist: Advertising tied directly to an AI answer might provide the cleanest argument publishers have had for getting paid. When a publisher’s work helps shape an answer, the revenue link has always been fuzzy; users typically subscribed long before they asked anything, and plenty of AI products run on free tiers. However, if your reporting powers an answer, and that answer generates money for the AI company through impressions or transactions, the chain from content to dollars suddenly looks a lot less abstract.

It’s also easier to measure than the old world ever was. Classic SEO involved a lot of educated guesswork based on keywords and clicks. AI queries are often longer and more specific, and the tooling is better at teasing out intent. That makes it much more feasible to identify which answers—and which underlying sources—most reliably push people toward a purchase.

OpenAI tried to tamp down commercialization anxiety by laying out its advertising first principles, including a promise that ads won’t change the substance of ChatGPT’s answers. In theory, if Coca-Cola buys a campaign, ChatGPT shouldn’t become any more (or less) likely to mention Coke than it was the day before. But it’s fair to wonder whether the system could still nudge users toward a transaction in general—say, buying a soft drink—while the ad sits nearby as the convenient button to press.

The new SEO: make it persuasive

Even if OpenAI keeps its side pristine, it can’t control how brands—and the publishers who want to be seen—adapt their own behavior. How effective those tactics will be is anyone’s guess, but it’s a safe bet they’ll be attempted. The emerging field of GEO (generative engine optimization) feels poised to grow a new limb: not only shaping whether you appear in an AI answer, but how strongly that answer prompts someone to act. You’re not just optimizing for presence, but also persuadability.

Right now, this is all mostly theory—and it’s entirely possible that Google, OpenAI, and everyone else will capture the ad upside for themselves. But the moment AI answers become meaningful revenue machines, marketers will obsess over which answers convert best, and what ingredients those answers contain. If publishers can demonstrate they supply the secret sauce, they’ll be in a stronger position to demand their slice.

Of course, “demonstrate” is doing a lot of work in that sentence. Real bargaining power requires proof, and proof starts with measurement—which won’t happen by accident. Figuring out how content appears in, influences, and performs inside AI answers is brand-new science, but it is science: testing, iterating, and using whatever levers exist—snippets, bot blocking, and dedicated GEO platforms among them.

Over the past 25 years, Silicon Valley slowly built tremendous platforms that ended up consuming the vast majority of advertising revenue, locking out the media in the process. And yes, AI could easily extend that same arc. But there’s an irony in monetizing AI answers with ads: It may end up creating the best opportunity for publishers to define exactly how much value they bring to them.

The post Ads in ChatGPT: a new revenue threat, or a new payout path for publishers? appeared first on The Media Copilot.

OpenAI is rolling out advertising in ChatGPT, but the dozens of publishers who signed content licensing deals with the company won’t see a cent of the ad revenue.

The company announced last week that ads will begin appearing for U.S. users on free accounts and the new $8/month ChatGPT Go tier. Paid Pro, Business and Enterprise subscriptions remain ad-free.

The Information reported that OpenAI has already pitched the placements to dozens of advertisers. The model is pay-per-view rather than pay-per-click, with ads appearing below ChatGPT’s responses — not within them.

The contrast with Perplexity is striking. The AI search startup launched its Publishers’ Program in 2024, offering revenue sharing when a publisher’s content is referenced in an ad-supported interaction. Perplexity later expanded this with Comet Plus, which pays publishers for traffic from its AI browser.

OpenAI has made no similar commitment. Publishers including The Atlantic, Vox Media, Axel Springer and others signed licensing deals that give OpenAI access to their content for model training and real-time retrieval. Those deals cover content access — not a share of downstream advertising revenue.

“Ads do not influence the answers ChatGPT gives you,” OpenAI wrote in its announcement. “We keep your conversations with ChatGPT private from advertisers, and we never sell your data.”

The move represents a reversal from CEO Sam Altman’s earlier stance. “Ads plus AI is sort of uniquely unsettling to me,” he said at a Harvard Business School talk in May 2024. “I kind of think of ads as a last resort for us.”

With over 800 million weekly active users, ChatGPT’s free tier represents significant monetization potential. For publishers watching their traffic decline as users get answers without clicking through, the lack of revenue sharing adds insult to injury.

The post OpenAI launches ChatGPT ads with no revenue share for publishers appeared first on The Media Copilot.

Behind the launch of OpenAI’s new training hub lies a tacit acknowledgment: AI is already reshaping journalism, whether newsrooms are ready or not.

Just before the holiday, OpenAI announced something new: the OpenAI Academy for News Organizations. An addition to OpenAI’s existing learning portal, the new Academy is described as a community where teams from media orgs—including editorial, business, product, and tech—can learn how to apply AI to their work, share insights, and maybe even work on the big, scary shifts that AI represents: declining traffic, AI’s factual fallibility, and the ongoing redefinition of knowledge work.

Reaction to the initiative was generally positive. Many commentators gave OpenAI credit, sometimes cautiously, for launching a journalism arm of its well-regarded training playground. That might be a surprise given that big chunks of the media world are suing OpenAI and its peers over alleged copyright violations. It probably helps that OpenAI turned to real journalists who’ve built programs and products that leverage AI at their orgs to lead the Academy.

That hands-on attitude carries over to content, too: Besides an AI Essentials for Journalists class (a nice primer for The Media Copilot’s extensive six-week AI for Journalists course 😉 ), there are resources and case studies from real newsrooms, like The Seattle Times Sales Prospecting Agent. Individual journalists can join the community, but there isn’t a forum, at least not yet.

OpenAI’s new journalism AI academy signals to me three things:

1. Newsrooms are interested in building, not dabbling. This is part of the systemic shift that I wrote about in my predictions for 2026: Applying AI at media organizations is shifting from dabbling with prompts to systemic tool deployments and audience-facing products. This is a difficult transition because it’s navigating in uncharted waters, and it requires building on the learnings of the past few years. Much of that is on offer in the Academy, and the OpenAI label will, in the minds of many, coat it with a veneer of legitimacy. That might push more newsrooms to accelerate building.
2. Custom GPTs remain an unsung feature of ChatGPT. Putting aside the concern that the Academy will inevitably be a ChatGPT lovefest (I have yet to find an article where Claude is the featured AI), several of the case studies feature Custom GPTs, sometimes prominently. I see that anecdotally as well. Even though OpenAI almost never talks about them, Custom GPTs have become key to workflows for many journalists, newsrooms, and knowledge workers more generally. OpenAI should tune to this obvious signal and support them better.
3. Don’t expect much guidance on the big questions. The Academy announcement said, “adopting new technology raises important questions for journalists and publishers, including concerns about trust, accuracy, and jobs. The Academy is built with those realities in mind.”
   
   Notice there’s no mention of intellectual property. I would not expect that many use cases from, say, TollBit or DataDome on how effective their bot blocking and monetization tools are. Or strategic guidance on how to treat AI systems as customers rather than a source of referrals. I fully expect the Academy to stay focused on how AI improves and alters ground-level journalism and newsroom operations, leaving out big-picture matters that might question OpenAI’s overall approach to content. That still includes a lot of important material, of course.

But the fact the Academy exists at all, and has good material, is perhaps a sign that we can tolerate that tension. Journalists can use ChatGPT and other AI tools to improve and accelerate their work while simultaneously understanding that the same technology is creating a massive shift in audience behavior that’s upending the economics of the entire industry. A nascent online Academy from just one of the frontier labs won’t magically solve those issues, but it’s a signal that solving them is in everyone’s interest.

The post The unstated message of OpenAI’s new journalist academy appeared first on The Media Copilot.

Adobe launched Photoshop, Adobe Express, and Acrobat inside ChatGPT on Tuesday, letting users edit photos, create designs, and manipulate PDFs using natural language prompts.

The integration is free and available globally on ChatGPT’s desktop, web, and iOS apps. Android support is coming soon for Photoshop and Acrobat, though Adobe Express already works on Android.

“Now hundreds of millions of people can edit with Photoshop simply by using their own words, right inside a platform that’s already part of their day-to-day,” David Wadhwani, president of digital media at Adobe, said in the company’s announcement.

Users can access the tools by typing the app name followed by an instruction. To blur a photo background, for example, you’d type: “Adobe Photoshop, help me blur the background of this image.”

The Photoshop integration allows users to adjust brightness, contrast, and exposure, apply effects like Glitch and Glow, and edit specific parts of images. Adobe Express offers access to design templates for invitations, social posts, and other content. Acrobat lets users edit PDFs, extract text and tables, merge files, and redact sensitive information.

Adobe said the launch builds on its work with “agentic AI” and the Model Context Protocol, an open standard for connecting AI assistants to external tools. The company previously introduced AI assistants for Photoshop and Adobe Express at its MAX conference.

For users wanting more control, Adobe said it’s “seamless to move from ChatGPT into Adobe’s native apps and pick up right where they left off.”

What this means for newsrooms: This could lower the barrier for journalists and editors who need quick image edits or PDF work but lack design training. The conversational interface removes the learning curve of Adobe’s professional tools. But the real question is whether the ChatGPT integration delivers results good enough for publication, or if it’s better suited for quick social media content and internal documents.

The post Adobe brings Photoshop, Express, and Acrobat to ChatGPT’s 800 million users appeared first on The Media Copilot.