Innovation Profs - 9/9/2025

Your weekly guide to generative AI tools and news

Author

Innovation Profs
September 09, 2025

Latest Gen AI News

Anthropic to pay authors $1.5 billion to settle lawsuit over pirated books used to train AI chatbots

Anthropic, the company behind the Claude family of language models, has agreed to a settlement of $1.5 billion dollars with authors who brought a class-action lawsuit against the company for training its models on pirated copies of their works. As part of the settlement, authors will receive about $3,000 per pirated work, with roughly 500,000 books covered in the settlement. The settlement, reportedly the largest copyright settlement to date and the first involving AI, was announced prior to judicial approval.

Judge skewers $1.5B Anthropic settlement with authors in pirated books case over AI training

But wait! Just yesterday, Judge William Alsup did not approve the settlement, leaving open the possibility that the Anthropic lawsuit could still go to trial. Alsup’s concerns seem to apply to both parties: How do we ensure that some of the affected authors don’t “get the shaft” by not being included in the settlement while also ensuring that Anthropic isn’t blindsided by future lawsuits “coming out of the woodwork”? A follow-up hearing is scheduled for September 25 that will discuss a potentially exhaustive list of the pirated works in question as well as more details about the claims process that authors will need to go through to receive their payment.

Warner Bros. Joins Studios’ AI Copyright Battle Against Midjourney

On a different legal front, Warner Bros is suing AI image generation platform Midjourney for copyright infringement, being represented by the same lawyers that are bringing a similar lawsuit on behalf of Disney and Universal. According to the lawsuit, “Without any consent or authorization by Warner Bros. Discovery, Midjourney brazenly dispenses Warner Bros. Discovery’s intellectual property as if it were its own.” The lawsuit specifically identifies Superman, Batman, Wonder Woman, Bugs Bunny, and Scooby-Doo as examples of Warner Bros. characters that have been reproduced by Midjourney.

Sign up for Gen AI Boot Camp

Gen AI Boot Camp: Join us in-person at Drake University or virtually for our next Generative AI Boot Camp. This one-day workshop will get you up to speed and using generative AI tools effectively in your work.

When: Friday, Sept. 12, 9 a.m. to 3 p.m.
Where: Drake University or on Zoom
Sign up here

Quick Hits

The Case for Using Small Language Models

Small Language Models (SLMs) are redefining enterprise AI by offering faster, more efficient, and cost-effective solutions compared to Large Language Models (LLMs).

hbr.org/2025/09/the-case-for-using-small-language-models

The art world is FINALLY cracking down on AI

Dragon Con's ejection of an AI art seller shows that the community is stepping up.

www.creativebloq.com/ai/ai-art/the-art-world-finally-drew-a-line-in-the-sand-over-ai-and-not-before-time

Why language models hallucinate

OpenAI’s new research explains why language models hallucinate. The findings show how improved evaluations can enhance AI reliability, honesty, and safety.

openai.com/index/why-language-models-hallucinate

These psychological tricks can get LLMs to respond to “forbidden” prompts

Study shows how patterns in LLM training data can lead to “parahuman” responses.

arstechnica.com/science/2025/09/these-psychological-tricks-can-get-llms-to-respond-to-forbidden-prompts

Switzerland releases an open-weight AI model

The model’s training data, model weights, and source code are open to the public.

www.theverge.com/ai-artificial-intelligence/770646/switzerland-ai-model-llm-open-apertus

The AI breakthrough that uses almost no power to create images

From creating art and writing code to drafting emails and designing new drugs, generative AI tools are becoming increasingly indispensable for both business and personal use. As demand increases, they will require even more computing power, memory and, therefore, energy. That's got scientists looking for ways to reduce their energy consumption.

techxplore.com/news/2025-08-ai-breakthrough-power-images.html

New AI model turns photos into explorable 3D worlds, with caveats

Openly available AI tool creates steerable 3D-like video, but requires serious GPU muscle.

arstechnica.com/ai/2025/09/new-ai-model-turns-photos-into-explorable-3d-worlds-with-caveats

Showrunner wants to use generative AI to recreate lost footage from an Orson Welles classic

The “Netflix of AI” next project is a reconstruction of Welles’ The Magnificent Ambersons.

www.theverge.com/entertainment/772635/showrunner-orson-welles-the-magnificent-ambersons

Tool of the week: Genspark AI Designer

Genspark AI Designer is a newly launched, agentic, no-code AI-powered design tool from Genspark. Genspark is a platform often described as an "all-in-one AI workspace". Released around August 2025, Genspark AI Designer positions itself as your personal AI “designer employee” that transforms a simple prompt into a full design suite instantly.

The tool creates complete brand systems, including logos and visual identities, packaging, websites, interior/exterior layout concepts, and marketing content like menus and posters.

Users can type in a description (“modern minimalist coffee brand”) and receive a full brand kit in minutes.

AI-generated image of the week

We’re still experimenting with Google’’s new Nano Banana image tool. One of the examples in the official guide to Nano Banana is a Kawaii-style sticker. So we made one for our beloved Drake Bulldogs. 

Prompt: A kawaii-style sticker of a bulldog wearing a blue "Drake" shirt

What we found

Google is rolling out new features to NotebookLM. Flashcards and quizzes will help test your knowledge on a topic. The revamped reports tool will let you specifying the structure, style and tone of your reports.