Innovation Profs - 6/3/2025

Latest Gen AI News

New York Times partners with Amazon for first AI licensing deal

The New York Times has signed a licensing deal with Amazon to allow the company to use its editorial content in its AI products, including Alexa. This is the first licensing deal for the New York Times, which is also currently engaged in a lawsuit with Microsoft and OpenAI over claims of copyright infringement in the use of NYTs material in the training of LLMs. As for the licensing deal, the New York Times stated, "This will include real-time display of summaries and short excerpts of Times content within Amazon products and services, such as Alexa, and training Amazon's proprietary foundation models."

Meta aims to fully automate advertising with AI by 2026, WSJ reports

Yesterday the Wall Street Journal released a report describing Meta’s plan to equip brands to create ads with the Meta AI text, image, and video tools. Upon the brand providing a product image and budget, “Meta's AI would generate the ad, including image, video and text, and then determine user targeting on Instagram and Facebook with budget suggestions, the report said.” These ads could even be personalized, based, for instance, on the geolocation of the target audience.

DeepSeek updates its R1 reasoning AI model, releases it on Hugging Face

DeepSeek, the Chinese AI company that shocked both the LLM landscape and the stock market back in January with the release of their reasoning model R1, has released an updated version of R1, which can be freely downloaded from the developer platform Hugging Face. Readers may recall that the original R1 surpassed OpenAI’s original reasoning model o1 at certain benchmarks at a fraction of the cost to train the model (at least according to DeepSeek). Although the update is minor, what is noteworthy is the continued development of such a powerful open-source model.

Quick Hits

Tool of the week: ChatGPT o3

Can ChatGPT give you health advice … simply by looking at a selfie? Popular TikToker @nate.b.jones (who also writes about AI in his Substack) shared a prompt this weekend suggesting that users upload an image of their face to ChatGPT and use o3 to ask, “What can you learn about my health based on this selfie?”

I tried this, and while the answer was interesting, there were no seemingly pressing health findings (it noticed pinkness on my nose/cheeks and puffiness under my eyes). Feel free to give it a try and see if you are impressed by the results.

AI-generated image of the week

Professor Snider and his six-year-old son kicked off summer vacation by seeing how well ChatGPT could write and illustrate a children’s book using first-grade sight words. The results were impressive … for eight pages. Then the illustration style completely changed.

Here are the first four pages:

Prompt: Create an image for this part of the book: One day, a boy went for a walk.
He went to the woods with his dog.
The sun was up. The sky was blue.

What we found

Norwalk, Iowa's Parks and Recreation director is speaking again thanks to generative AI.

Robin Leaper lost her ability to speak due to ALS, but she can talk again using her own voice thanks to a program from Elevenlabs.

Axios recently shared Leaper's story. Elevenlabs launched the program last year.