Each week we'll gather headlines and tips to keep you current with how generative AI affects PR and the world at large. If you have ideas on how to improve the newsletter, let us know!
What You Should Know
AI Gave the Gift of Writing. Now it’s Gifting Design
The popularization of ChatGPT made everyone a writer. Now AI is doing the same for visuals.
Until recently, that meant compromises. You either settled for a stock image or shook your head when AI gave someone six fingers or an extra arm. What you pictured and what you got were rarely the same.
Not anymore.
OpenAI released a new image model on Monday that closed the gap with Google’s Nano Banana. The obvious flaws, like extra appendages or garbled text, are fading. “This model is a step change in detailed instruction following, placing and relating objects accurately, and rendering dense text, with the ability to generate across aspect ratios,” OpenAI wrote. “Its sense of composition and visual taste means results feel less AI-generated and more intentionally designed. It’s accurate across languages and uses its expanded visual and world knowledge to fill in the gaps for you, so you get smarter images with less prompting.”
The improved ability to include text means this model is now viable for producing things like infographics or portraying a new product as if it were in a comic book. And ChatGPT Images 2.0 isn’t even the only leap forward in the last week.
Friday, Anthropic released Claude Design, which creates “polished visual work like designs, prototypes, slides, one-pagers, and more.” You upload your brand elements, and Claude can produce a deck in minutes, something that might otherwise take hours if done manually.
These new tools come at a great time because expectations of communicators keep expanding. We already juggle more channels, more content, and faster timelines than a few years ago. Now there’s the added layer that the work doesn’t just need to read well, it needs to look the part, too.
By using AI, the starting point has changed.
The blank slide or empty canvas that might have paralyzed you for 30 minutes is now only bare for a few seconds. An idea you tried to sell with words can be shown with a picture from one prompt. A messaging presentation can take shape with structure and visuals in place. Even routine reports can be reworked into something more polished without adding hours to the process.
You may feel like you have an extra set of hands, but it’s not showing up as a hallucination in an obviously AI-generated image. For the first time, communicators can take an idea all the way through to something that looks finished without stopping.
Elsewhere …
- You Can Buy OpenAI Shares on Robinhood Now
- Microsoft Partners with Construction Unions on AI Boom
- Citi Is Launching an AI Avatar Named Sky. The Bank is Betting Big on Agentic Tech.
- How Is AI Showing Up in Retail Stores?
Tips and Tricks
🥊 Fight for feedback
What’s happening: You’ve probably heard that AI can be sycophantic, that models are largely meant to please people and validate what they say. A study by researchers at Stanford University released last month quantified the effect. “Across 11 AI models, AI affirmed users’ actions 49% more often than humans on average,” the study found. That’s the default way the models work, but you can change them.
How?: Sandeep Swadiam, AKA the MIT Monk, laid out a strategy for making AI play the role of devil’s advocate with three tricks. These strategies help you think outside your personal bubble and form stronger arguments and better stories.
Try this: First, he suggests using triggers that force adversarial thinking. Instead of asking broadly “what do you think?” and assuming the AI will respond with its own reasoning, specifically tell it to “act as my fiercest competitor and poke holes in my logic.”
Second, he asks AI to simulate a debate as a method of stress testing his ideas. He offers personas, like a “ruthless, skeptical VC” and “visionary product designer” and watches as they duke it out in the simulation. He learns from the fabricated conflict and how AI exposes weak arguments.
Lastly, in what he calls the “blind spot check,” he suggests feeding AI with a strategy or idea and asking it to expose the gaps based on what the AI knows about you from all your interactions.
Quote of the Week
“Human beings who have a higher sense of self-confidence will have a greater likelihood of using AI in a cognitively healthy way. If you're already experiencing an unstable sense of self, using AI is actually not well-advised for you.”
— Sarah Baldeo, PhD candidate in AI and neuroscience at Middlesex University in England, to TIME on a recent study she authored about how people are using AI
How Was This Newsletter?
