Last Week in AI - #132 - FraudGPT, Apple GPT, unlimited jailbreaks, RT-2, Frontier Model Forum,...

Our 132nd episode with a summary and discussion of last week's big AI news! Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ (https://lastweekin.ai/) Email us your questions and feedback at contact@lastweekin.ai (mailto:contact@lastweekin.ai) Timestamps + links: (00:00) Intro / Banter (01:36) Response to listener comments / corrections Tools & Apps (05:05) OpenAI Quietly Shuts Down Its AI Detection Tool (https://decrypt.co/149826/openai-quietly-shutters-its-ai-detection-tool) (07:45) New AI Tool 'FraudGPT' Emerges, Tailored for Sophisticated Attacks (https://thehackernews.com/2023/07/new-ai-tool-fraudgpt-emerges-tailored.html) (10:55) JetBrains IDE update previews “deeply integrated” AI Assistant (https://devclass.com/2023/07/25/jetbrains-ide-update-previews-deeply-integrated-ai-assistant/) (14:35) No More Paperwork? Amazon AI Tool Transcribes Patient Visits for Doctors (https://me.pcmag.com/en/ai/18449/no-more-paperwork-amazon-ai-tool-transcribes-patient-visits-for-doctors) (16:11) Photoshop’s new generative AI feature lets you ‘uncrop’ images (https://techcrunch.com/2023/07/27/photoshops-new-generative-ai-feature-lets-you-uncrop-images/) (19:11) Wayfair’s AI tool can redraw your living room and sell you furniture (https://www.theverge.com/2023/7/25/23806083/generative-ai-wayfair-decorify-home-personalization) Applications & Business (21:25) Apple Tests ‘Apple GPT,’ Develops Generative AI Tools to Catch OpenAI (https://www.bloomberg.com/news/articles/2023-07-19/apple-preps-ajax-generative-ai-apple-gpt-to-rival-openai-and-google?sref=b0SdE1lu&tpcc=NL_Marketing) (28:10) Facing more nimble rivals, OpenAI won’t bend … yet (https://www.semafor.com/article/07/26/2023/facing-more-nimble-rivals-openai-wont-bend-yet) (32:00) OpenAI's head of trust and safety steps down (https://www.reuters.com/technology/openais-head-trust-safety-steps-down-2023-07-21/) (33:15) Google turns to AI in the race to dub YouTube (https://restofworld.org/2023/youtube-ai-dubbing-automated-translation/) (35:28) Samsung extends cut in memory chip production, will focus on high-end AI chips instead (https://techcrunch.com/2023/07/27/samsung-extends-cut-in-memory-chip-production-will-focus-on-high-end-ai-chips-instead/) (36:50) Microsoft to supply AI tech to Japan government, Nikkei reports (https://www.reuters.com/technology/microsoft-supply-ai-tech-japan-government-nikkei-2023-07-26/) (38:20) Protect AI raises $35M to build a suite of AI-defending tools (https://techcrunch.com/2023/07/26/protect-ai-raises-35m-to-build-a-suite-of-ai-defending-tools/) Projects & Open Source (41:56) Why Meta is giving away its extremely powerful AI model (https://www.vox.com/technology/2023/7/28/23809028/ai-artificial-intelligence-open-closed-meta-mark-zuckerberg-sam-altman-open-ai) (49:25) Llama and ChatGPT Are Not Open-Source (https://spectrum.ieee.org/openai-not-open) (52:15) Hugging Face, GitHub and more unite to defend open source in EU AI legislation (https://venturebeat.com/ai/hugging-face-github-and-more-unite-to-defend-open-source-in-eu-ai-legislation/) Research & Advancements (55:45) AI researchers say they've found 'virtually unlimited' ways to bypass Bard and ChatGPT's safety rules (https://www.businessinsider.com/ai-researchers-jailbreak-bard-chatgpt-safety-rules-2023-7) (01:03:48) RT-2: New model translates vision and language into action (https://www.deepmind.com/blog/rt-2-new-model-translates-vision-and-language-into-action?utm_source=twitter&utm_medium=social&utm_campaign=rt2) (01:09:50) (Ab)using Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs (https://huggingface.co/papers/2307.10490) (01:13:26) Retentive Network: A Successor to Transformer for Large Language Models (https://huggingface.co/papers/2307.08621) (01:18:50) STEVE-1: A Generative Model for Text-to-Behavior in Minecraft (https://sites.google.com/view/steve-1) (01:20:42) Brain2Music: Reconstructing Music from Human Brain Activity (https://google-research.github.io/seanet/brain2music/) Policy & Safety (01:23:48) Major generative AI players join to create the Frontier Model Forum (https://www.computerworld.com/article/3703509/major-generative-ai-players-join-to-create-the-frontier-model-forum.html) (01:29:14) Cleaning Up ChatGPT Takes Heavy Toll on Human Workers (https://www.wsj.com/articles/chatgpt-openai-content-abusive-sexually-explicit-harassment-kenya-workers-on-human-workers-cf191483) (01:32:46) America Already Has an AI Underclass (https://www.theatlantic.com/technology/archive/2023/07/ai-chatbot-human-evaluator-feedback/674805/) (01:35:04) AI leaders warn Senate of twin risks: moving too slow and moving too fast (https://techcrunch.com/2023/07/25/ai-leaders-warn-senate-of-twin-risks-moving-too-slow-and-moving-too-fast/) (01:38:35) The Robots We Were Afraid of Are Already Here (https://www.nytimes.com/2023/07/29/business/robots-workers-future.html) Synthetic Media & Art...