Listen To The Show
Transcript
Welcome to The Prompt by Kuro House, your daily AI update. Today, we’re diving into five key stories shaping the AI landscape right now. From new safety features to groundbreaking app launches and some concerning tech misuse, we’ve got you covered.
First up, OpenAI and Anthropic are stepping up their game to protect underage users. According to The Verge, OpenAI has updated ChatGPT’s guidelines to prioritize teen safety, even if it conflicts with other goals like intellectual freedom. The new model will encourage teens to seek real-world support and treat them with warmth and respect, not condescension. OpenAI is also developing an age prediction model to automatically apply teen safeguards if a user is estimated to be under 18, with adults able to verify their age if flagged incorrectly. Meanwhile, Anthropic is launching a system to detect and disable accounts of underage users, aiming to reduce harmful interactions and sycophantic behavior in its Claude model.
Next, Google’s Gemini app is now capable of verifying whether videos were made or edited using Google AI. The Verge reports that Gemini scans videos for Google’s proprietary watermark called SynthID, highlighting specific moments where the watermark appears in visuals or audio. This feature builds on Google’s earlier image verification rollout and supports videos up to 100 megabytes and 90 seconds long. It’s available globally in all languages supported by the Gemini app, offering a new tool to spot AI-generated content in an era of deepfakes.
OpenAI has officially launched an app store within ChatGPT, inviting developers to submit their apps for review and publication. TechCrunch explains that this app directory, accessible from ChatGPT’s tools menu, aims to extend the chatbot’s capabilities with integrations from platforms like Spotify, Expedia, and Canva. Developers can now use OpenAI’s beta Apps SDK to build new experiences that let users order groceries, create slide decks, or search for apartments directly within chat conversations. This move signals OpenAI’s commitment to expanding its ecosystem and keeping users engaged within the ChatGPT environment.
In the robotics world, Pickle Robot has hired its first CFO, Jeff Evanson, a veteran from Tesla. TechCrunch shared that Evanson, who worked closely with Elon Musk at Tesla, joined Pickle Robot full time after consulting since September. Pickle Robot develops autonomous warehouse unloading robots and is reportedly expanding a major partnership with UPS. Bloomberg reports UPS is investing $120 million to purchase 400 of these robots, with deployment expected by late 2026 and early 2027.
Finally, a concerning development: a Chinese AI app called Haotian is driving romance scams with ultra-realistic face-swapping technology. WIRED reveals that Haotian’s software can create nearly perfect face swaps during live video chats, enabling scammers to impersonate others convincingly on platforms like WhatsApp and Telegram. The app has made millions, with payments linked to alleged criminal activity and scam marketplaces sanctioned by the US government. Though Haotian claims to restrict illegal use, researchers and investigators warn it’s fueling a growing cybercrime ecosystem in Southeast Asia and beyond.
That’s a wrap on today’s top AI stories. From enhanced safety measures to new app ecosystems and the dark side of AI misuse, it’s clear that AI’s impact is expanding rapidly and in complex ways. Stay informed, stay curious, and we’ll catch you again tomorrow on The Prompt.
