Innovation Profs - 9/5/2023

Generative AI News

OpenAI launches ChatGPT Enterprise, the company’s biggest announcement since ChatGPT’s debut

Early last week, OpenAI rolled out ChatGPT Enterprise, which includes unrestricted access to GPT-4, high-speed performance, and API credits. Another key feature is that ChatGPT Enterprise will allow companies to fine tune ChatGPT on their own data. A lower tier of usage for small businesses, ChatGPT Business, will be rolled out at a later date.

OpenAI disputes authors’ claims that every ChatGPT response is a derivative work

OpenAI recently responded to two class action lawsuits, each filed by a team of book authors (one of which includes Sarah Silverman). According to the lawsuits, ChatGPT’s training set included pirated copies of the authors’ works, which, they claim, constitutes copyright infringement. In seeking to dismiss all but one of the claims filed against them, OpenAI argues that the authors misconstrue the scope and purpose of copyright and, notably, that the patterns of syntax and word frequencies occurring in an author’s book are not subject to copyright.

US Copyright Office opens inquiry into laws needed to rein in genAI

On the topic of copyright, the US Copyright Office has issued a notice of inquiry on copyright and generative AI. In particular, the agency has been fielding feedback on “whether legislative or regulatory steps are warranted.” To date, four public listening sessions and two webinars have been held, and comments can still be submitted here through October 18.

AI is reshaping the workplace – but what does it mean for the health and well-being of workers?

What are some of the possible health impacts of AI for workers? Although AI might enhance worker safety by performing riskier tasks or identifying workplace hazards, according to the author, challenges included the possibility of inequities in the labor market and increased stress as workers strive to keep up with machines. Ultimately, more work remains to understand the effect of AI on the human workforce.

Google made a watermark for AI images that you can’t edit out

Early last week, Google released a new tool, developed by researchers at Google DeepMind, that watermarks AI-generated images in a way that is easy to detect using DeepMind’s tool but undetectable to the human eye. DeepMind’s CEO Demis Hassabis hopes that the tool might eventually be used for an internet-wide standard, an important step to curb the spread of deepfake technology. Other companies are pursuing similar strategies to identify AI-generated image.

Know someone who wants to learn more about generative AI? Forward them this newsletter or send them this link to sign up.

Tool of the week: Kaiber.ai

Motion by Kaiber transforms your static text into dynamic animations. Type in text or upload an image, and Kaiber will turn it into video.

AI-generated image of the week

It’s Cy-Hawk week here in the great state of Iowa. This left us wondering how well Midjourney could create the team mascots. We’re still not so sure…

Prompt:  university of iowa mascot herky the hawkeye battles with iowa state mascot cy the cardinal on a football field of a packed stadium, photo, 4k, realistic --ar 16:9 --v 5.2

Generative AI tip of the week

Learn to speak generative AI: CNET published a list of AI terms everyone should know. Check out the list - from AI Safety to Zero-Shot Learning.

Get starting with Generative AI

New to generative AI? Here are some places to start…

• NightCafe image prompting tips

• Learn ChatGPT prompting

• Best AI image generators

What we found

Dwayne “The Rock” Johnson - Disney princess? Instagram account @the_ai_dreams used generative AI to reimagine The Rock as Disney princesses. Our favorite? Elsa.

Bonus: The account recently posted Jack Black as Disney princesses.