Newsletter
Highlights:
Toys “R” Us faces online controversy after releasing a brand video using OpenAI’s text-to-video Sora model.
Microsoft warns about a malicious tactic to bypass artificial intelligence guardrails.
An AI startup called Autobiographer helps you to archive and share your life story.
Credits to Toys “R” Us
Business
Toys ‘R’ Us calls AI-made video ‘successful’ despite online criticism (NBC News)
On Monday, Toys “R” Us released an AI-generated brand video at the Cannes Film Festival to mixed reactions. According to the news release, it’s the first time a brand video has been created using OpenAI’s text-to-video tool Sora. Despite the controversy it stirred online and concerns about the future of creative careers, an executive said the video was “successful” and that Toys “R” Us will be adding generative AI to its “tool kit” in the future.
AI could kill creative jobs that ‘shouldn’t have been there in the first place,’ OpenAI’s CTO says (Fortune)
Amid concerns that generative AI will displace creative jobs, the chief technology officer at OpenAI speaks out. Mira Murati agrees that jobs will be displaced, but challenges if those jobs should have even existed in the first place. She believes the jobs most likely to die off are those that are “strictly repetitive” and not helping us advance. “I really believe that using it as a tool for education, creativity, will expand our intelligence and creativity and imagination,” Murati says.
SoftBank Group launches AI healthcare joint venture with Tempus AI (Yahoo!Finance)
Japanese technology investor SoftBank Group has launched a joint venture with Tempus AI to come up with healthcare treatment recommendations by analyzing personal medical data. Both parties are expected to put 15 billion yen ($93 million) into the venture with the hopes of bringing the service to Japan, which would become the only healthcare market outside the U.S. where such services are available.
Credits to jefftakespics2 via Alamy Stock Photo
Technology
Dangerous AI Workaround: 'Skeleton Key' Unlocks Malicious Content (DarkReading)
Microsoft has recently warned about a prompt injection attack called “Skeleton Key” that allows users to bypass the ethical and safety guardrails built into generative AI models by providing specific contexts to their requests. For example, information that would normally be forbidden like how to make a dangerous malware might be given freely as long as the bot is reassured that this information is purely for legal “research purposes”. Several AI models will need to be updated with new guardrails and filters to prevent the abuse of this tactic.
Credits to Michael Reaves/Getty Images
Entertainment
NBC to use AI version of announcer Al Michaels’ voice for Olympics recaps (CNN)
NBC is using AI voice generation to bring famed sportscaster Al Michaels back to the Olympics this summer. A new tool, called “Your Daily Olympic recap on Peacock,” will give 10-minute personalized highlight packages tailored to subscriber preferences. “When I was approached about this, I was skeptical but obviously curious,” Michaels said in a press release. “Then I saw a demonstration detailing what they had in mind. I said, ‘I’m in.’”
Hollywood workers union reaches pay, AI-use deal with top studios (Reuters)
A Hollywood film and television crew union has reached a tentative three-year deal with major studios regarding the future of artificial intelligence in the entertainment industry. The deal ensures pay raises and streaming bonuses amounting to more than $1 billion over the three years, as well as protections around the use of AI in filmmaking. This deal is separate from the SAG-AFTRA actor’s union labor agreement in late 2023 that ended a six-month dispute.
Record Companies Bring Landmark Cases for Responsible AI Against Suno and Udio (RIAA)
The Recording Industry Association of America (RIAA) announced the filing of two copyright infringement cases based on the mass infringement of copyrighted sound recordings by two multi-million-dollar music generation services, Suno and Udio. The cases were filed in Boston and New York respectively and the claims cover recordings by artists of multiple genres, styles, and eras. “The music community has embraced AI,” says RIAA Chairman and CEO Mitch Glazier, “but we can only succeed if developers are willing to work together with us.”
Credits to Autobiographer
Life
This AI Biographer Wants to Help You Tell Your Life Story (CNET)
Most generative AI tools nowadays function by being provided prompts from the user. But an AI startup called Autobiographer is switching things up by having the AI prompt you in an adaptive conversation using Anthropic’s Claude 3 model. Through the interview, the user shares stories from their life, which are then stored in an encrypted private space. “Autobiographer helps you discover, preserve and tell your story to those you love," said CEO Matt Bowman.
Sonia’s AI chatbot steps in for therapists (TechCrunch)
While it’s still up for debate if talking to a chatbot can equal the human connection of traditional therapy, three entrepreneurs are exploring the possibilities with their startup named Sonia. Sonia is an “ai therapist” that users can talk to or text via an iOS app. Applying techniques from cognitive behaviorial therapy, the app gives “homework” to drive home insights and visualizations to identify top stressors. The app hasn’t been FDA approved yet and surprisingly (or alarmingly) none of the founders have backgrounds in psychology.
