⚡ OpenAI Brings Ads to ChatGPT After Years of Saying No

ChatGPT is about to look a little different for some users. OpenAI plans to start testing ads for free U.S. users and its low-cost Go plan, marking a shift after years of pushing an ad-free vision. Paid tiers will stay ad-free.. for now. Let’s dive in!

In today’s insights:

LATEST DEVELOPMENTS

Evolving AI: OpenAI will start testing ads in ChatGPT for free U.S. users soon.

OpenAI will test ads inside ChatGPT in the U.S. within weeks. Adult free users and the low-cost Go plan will see ads, with paid tiers staying ad-free. Ads sit below answers, carry clear labels, and avoid topics like politics or health. OpenAI says replies are not shaped by ads and user data will not be sold. Users can review why an ad appears and share feedback.

Earlier, OpenAI leaders publicly pushed back on ads, with Sam Altman often framing an ad-free ChatGPT as key to trust. That stance slowly softened over the past year, as costs rose and Altman hinted ads could arrive one day, setting the stage for this shift.

Ads in ChatGPT is OpenAI turning the chatbot into a new kind of “intent surface”: when someone asks what to buy, which tool to use, or how to solve a problem, the next step can be a sponsored link right under the answer. For free U.S. users (and the $8 Go tier), that means you’ll start weighing “helpful” vs “paid” in the same scroll, with controls to see why you got an ad and to dismiss it, plus guardrails around politics and health. What makes this more interesting is that Altman spent years pushing back on ads, warning they could hurt trust, and now OpenAI is leaning into them anyway.

Evolving AI: The question isn’t if you’ll deploy voice AI. It’s how.

This handbook from ElevenLabs explores the modern approaches to building scalable, humanlike voice agents from full in-house builds to hybrid orchestration frameworks.

Learn from real-world enterprise examples and find the model that fits your business and technical goals.

Free Download →

Evolving AI: A Stanford-led study links AI use to higher wages and lower inequality.

Researchers from Stanford and the Barcelona School of Economics studied how generative AI changes tasks, skills, and job choices. Their model tracks how workers build skills and move between occupations as AI reshapes work through automation, augmentation, and a third channel called simplification. That simplification lowers skill barriers for many tasks, raising overall pay and reducing inequality. Some high-end roles like engineers and executives may see pay drops, even as average wages climb.

If Althoff and Reichardt’s model is even directionally right, the big shift isn’t “AI replaces jobs,” it’s “AI reshapes skill ladders.” Their result comes from “simplification” making harder tasks doable for more people, which pushes wages up on average and squeezes wage gaps. Workplaces can still land in messier territory, with some groups losing out and inequality moving either way depending on adoption choices and bargaining power. The play for companies is clear: redesign workflows so junior talent can execute higher-value steps with AI guardrails, then update pay bands, training, and promotion paths before wage compression and role churn hit your org by surprise.

Evolving AI: Greenland’s minerals are back in focus as AI and chip supply chains meet geopolitics.

Greenland hosts major deposits of rare earths, gallium, and germanium used in magnets, semiconductors, fiber optics, and data centers. Political interest from the U.S. ties these materials to AI growth and security aims. Yet extraction is slow, with few active mines, high costs, harsh conditions, and price swings limiting progress. Experts say refining capacity, more than raw ore, decides real value.

Greenland sounds like a shortcut to AI hardware independence, but ore in the ground doesn’t fix supply risk. Rare earth value shows up at the processing and magnet stage, and the U.S. is still building that base, which is why the DoD’s long-term bet on MP Materials is such a big signal. Greenland mining would still need ports, roads, power, trained crews, plus political and legal clarity after fights like Kvanefjeld. Near-term, the faster win is refining capacity, not new territory.

Dictate full-context prompts and paste clean, structured input into ChatGPT or Claude. Wispr Flow preserves your nuance so AI gives better answers the first time. Try Wispr Flow for AI.

See a demo

QUICK HITS

💰 Musk seeks up to $134 billion from OpenAI and Microsoft.

🤝 South Korea's Lee, Italy's Meloni agree to strengthen cooperation in AI, chips.

🇸🇪 Song banned from Swedish charts for being AI creation.

⛔ China blocks Nvidia H200 AI chips that US government cleared for export.

🚗 AI system aims to detect roadway hazards for TxDOT.

 *partner link