Plus: OpenAI revives its robotics team.
Welcome back to The Prompt.
Today, a group of 13 current and former employees at both OpenAI and Google Deepmind published an open letter warning that leading AI companies are not doing enough to make their AI systems safe and accusing them of prioritizing profits and growth instead. The letter highlights the lack of transparency from companies about the risks and limitations of their AI applications, which could “potentially result in human extinction.” The letter also called for more protections for whistleblowers by developing processes for raising concerns to company decision makers and removing restrictive non-disparagement agreements.
This letter is just one more battle line being drawn in the AI industry, where bitter arguments over the right balance of safety and innovation are pitting billionaire versus billionaire, as we’ll discuss later in this edition.
Now, let’s get into the headlines.ETHICS + LAW
The Ansel Adams estate publicly chastised Adobe for selling AI-generated stock images in the style of the landscape photographer best known for his black-and-white images of the American West. “You are officially on our last nerve with this behavior,” the estate posted on Threads, adding that it has been trying to reach Adobe since August 2023 to take down the images. In response, Adobe said it removed the images, which violated its AI content policies.
Meanwhile, an AI-generated image with the slogan “All eyes on Rafah,” went viral last week with over 50 million shares on Instagram and Facebook. The image was created to draw attention to Israel’s airstrike on an encampment for the displaced in Gaza — but it also spurred an onslaught of AI-generated spinoff images showing distressed Palestinian women and children, according to 404. The original image was created by a Malaysian woman, but the version that went viral was modified by another Malaysian artist, who removed a watermark that signaled the image was made with AI.TALENT SHUFFLE
OpenAI is reviving its previously abandoned in-house robotics team, multiple sources told Forbes. The company is hiring research engineers for a team that will train and integrate AI models into robots manufactured by “external partners” like Figure AI. Before it released ChatGPT, OpenAI had considered building a “general purpose robot,” but dropped the effort in late 2020 due to a lack of training data.CHIP WARS
On Sunday, Nvidia announced its next AI chip, dubbed Rubin, which will be available in 2026.
