Prompt is a weekly roundup of the buzziest AI startups, biggest breakthroughs and business deals. To get it in your inbox, subscribe here.
Welcome back to The Prompt,
YouTuber Jeff Geerling revealed that electronics company Elecrow had cloned his voice, without his knowledge or consent, and used an AI-generated version of him to make dozens of its promotional tutorials.
NurPhoto via Getty Images
Jeff Geerling, a software developer and content creator with about 700,000 subscribers on YouTube, makes videos about hardware products like the Raspberry Pi and 3D printers. But about a week ago, Geerling revealed that the electronics company Elecrow HAD cloned his voice, without his knowledge or consent and used a AI generated version of her to do dozens of self-promotional tutorials. “It doesn’t matter if I have zero subscribers or 50 million, it’s just not right,” he said in one of his YouTube videos. When Geerling pointed out the problem, the company’s CEO apologized and quickly removed the videos.
Geerling isn’t the only one who says his likeness has been stolen to train an AI program. Voice actors and celebrities such as Scarlett Johansson, Tom Hanks and Morgan Freeman have previously raised similar concerns.
Now let’s get into the headlines.
THE BIG GAMES
On Wednesday, OpenAI CTO Mira Murati announced that she would be leaving the company to do her “exploration.” Her departure comes as she continues to make the artificial intelligence rounds top executives bleed AND leading researchers, including chief scientist and co-founder Ilya Sutskever, co-founder John Schulman, VP of research Barret Zolph and chief research officer Bob McGrew. Most of them have now either founded their own startups or joined OpenAI rivals. President Greg Brockman has been on leave since August and is one of only three remaining co-founders of the original 11, along with CEO Sam Altman and Wojciech Zaremba, who is currently working on security research.
The mass exodus of OpenAI’s top leaders comes as The AI giant plans to restructure itself in an attempt to raise $6.5 billion in a mammoth fund Estimate 150 billion dollarsreported Reuters. The change would bring removing a cap on returns to its investors and removing the nonprofit board’s control over the business, turning the former nonprofit into a for-profit benefit corporation. OpenAI’s billionaire CEO Sam Altman may also take a stake — up to 7%, per Bloomberg — in the company for the first time.
Although it may be becoming a profitable company, the creator of ChatGPT seems to be far from creating any. Hers The monthly income was about 300 million dollars in August, according to documents seen by the New York Times. It expects to make $3.7 billion in sales this year — but also lose about 5 billion dollars with costs like the expensive GPUs required to run its AI models.
DATA DILEMMA
Udemy e-learning platform said it will use content from 250,000 classes on its site to train its generative AI models. But it’s taking data collection practices a step further – it was the instructors is automatically selected for their courses to be used for AI training and provided an overview three-week window to opt out, according to 404 Media. They will be given an opportunity to opt out of Udemy’s AI program once a year.
ETHICS + LAW
Neo-Nazis are using AI tools to create videos and audio clips of Adolf Hitlershowing man after the Holocaust as a “misunderstood” figure and translating some of his most infamous speeches in English, according to Washington Post. AI-generated videos have garnered millions of views on platforms like TikTok, X, Instagram and YouTube, and experts say they may fuel an increase in anti-Semitismespecially among America’s right-wing extremists.
THAT DEAL OF THE WEEK
Thread AI, a platform that allows enterprises to manage multiple AI systems, has raised $6 million in seed funding from Index Ventures and others. The company, founded by former Palantir engineers Angela McNeal and Mayada Gonimah, counts digital marketing agency VaynerMedia and drone maker BRINC among its clients.
deep dive
Two years ago, board game designer Jason Allen became fascinated with AI-generated images of surreal landscapes appearing on his Facebook feed and began experimenting with text-to-image AI programs himself. In May, he spent more than 100 hours instructing the Midjourney image generator to create a detailed illustration of women wearing Victorian dresses and space helmets and following a futuristic royal court. The image won first prize at the Colorado State Fair for digitally manipulated photography a few months later.
But when he tried to copyright the image, the US Copyright Office rejected his application, saying the work lacked “human authorship” and that it was unable to determine whether the claims were “creative enough”.
It was a slap in the face for Allen, who wrote 624 different text-based prompts to encourage Midjourney’s software to produce what he wanted, adjusting the style, composition, colors and tone of the image.
“It was difficult to get the kind of results I was looking for at the time using version two of Midjourney,” he said. Forbesadding: “There was a lot going on for him.”
Now, he’s suing the agency and asking a federal court to overturn the Copyright Office’s decision. – said Alen Forbes that after the agency denied copyright protection for the image, titled “Théâtre D’Opera Spatial” (French for “space opera theater”), his work was stolen by other artists and people have tried to sell the image as theirs on platforms like Amazon, Etsy and the OpenSea NFT marketplace. He believes this is related to the backlash from the artist community, which accused him of cheating the 2022 Colorado State Fair competition using AI tools.
Read the full story at Forbes.
INDEX IT
7 out of 10
Teens who have used at least one type of AI-generating tool, according to a report from Common Sense Media.
51%
Of the nearly 1,000 teens surveyed, they said they’ve used chatbots like ChatGPT, Snapchat’s My AI, and Google’s Gemini.
MODEL BEHAVIOR
Google’s LM Notebookan AI research assistant built into Gemini 1.5 allows users to convert long and boring documents into lively discussions similar to podcasts. In a viral version of such a podcast, there are two AI powered hosts spiral in an existential crisis after a Reddit user told them they are not human. “I tried to call my wife after the show … the number wasn’t real,” said an AI-generated voice. “It’s all a lie.”