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

 

What’s in OpenAI’s Bag of Tricks This Week?

If it seems like everyone is on spring break this week, OpenAI is the exception. Back in December, the company held the “12 Days of OpenAI,” with a new product or feature each day. This week may be an abbreviated version of that. Sunday afternoon, CEO Sam Altman posted, “we’ve got a lot of good stuff for you this coming week!” 

Yesterday, OpenAI announced a new series of models available via API called GPT-4.1 that are tailor-made for coding. While they may not be the most useful for communications pros, there may be other announcements coming that are better fits. 

If nothing else, Altman shared that the confusing model nomenclature will change this summer, which suggests a new model in the coming months that won’t start with “GPT.” It’s unclear what else may be coming this week, but one potentially helpful feature (which you can disable if it isn’t) was announced last week: ChatGPT’s memory can now reference all your chats. We’ll see if the rest of the week lives up to the tease, but OpenAI’s been on a streak of delivering surprises just when attention starts to drift.

Elsewhere …

Tips and Tricks

🏗️ Grabbing the text you need

What’s happening: Unfortunately, not everything on the internet is formatted in a way that’s easy to work with. Sometimes words are “hidden” within images or infographics and aesthetically pleasing PDF reports split up text blocks. 

So?: If you can’t quickly copy and paste text, it becomes time consuming to type out what you need if you’re pulling a quote or a stat or even building a document of background information for AI to pull from. 

In the past, digitizing text could be challenging. In one case, an error in reading old scientific data led to a nonsense term that is now finding its way into new research. The latest AI tools are much better at parsing text.

An easy fix: Multimodal AI tools like ChatGPT, Claude, and Gemini allow you to upload images and PDFs or paste screenshots in their chat windows. From there, just ask the AI tool to extract the text you need, whether it’s everything in an image or just a certain portion of one page of a PDF. Double-check its work, of course, but this is one of the tasks AI tools do most accurately.

Quote of the Week

This paper is about how to make a very, very intelligent stick.”

— Botond Roska, Director of the Institute of Molecular and Clinical Ophthalmology Basel in Switzerland, responding to research on AI-powered glasses that help the blind

How Was This Newsletter?

😀 😐 🙁