The ChatGPT Observer EP23

The ChatGPT Observer EP23

If you're short on time, here's your one-minute summary of this week's key AI news and advancements:

?? Key Statistics: Washington Post analysis reveals 20% of AI chatbot interactions involve creative writing tasks, 17%+ are for homework assistance, and 15% are work-related.

?? This Week's Spotlight: Google NotebookLM reimagines note-taking with AI-powered summarisation and idea generation capabilities.

?? ChatGPT & OpenAI News: OpenAI introduces Structured Outputs for developers, welcomes AI safety expert Zico Kolter to its board, and Sam Altman teases mysterious "Project Strawberry".

? Claude Sonnet News: Anthropic launches an AI safety bug bounty program to identify vulnerabilities in its AI models.

?? Humanoid Robots: Figure 02 is ready for work, while Nvidia enables control of humanoid robots using Apple Vision Pro.

?? Corporate AI Developments: ByteDance launches an AI text-to-video tool, Zoom unveils AI-powered virtual backgrounds, and Google announces major updates for Gemini 1.5 Flash.

?? AI Innovations: Chinese researchers develop the world's first light-powered AI training system, and AI accelerates drug development for mental health disorders.

?? AI in Business: A judge uses ChatGPT to assist in reaching a verdict, sparking debate over AI's role in the courtroom.

? Responsible AI: Tech firms seek more time for EU AI Act compliance, while X pauses EU data processing for AI training.

?? For more details, read the full newsletter!


?? Key Statistics: How People Really Use AI Chatbots

A recent Washington Post analysis of nearly 200,000 English-language conversations with AI chatbots has revealed fascinating insights into how people are utilising this technology in the real world. Here are the key statistics that paint a picture of current chatbot usage:


Source: Washington Post / WildChat


  1. Creative Endeavours (20%): One-fifth of all requests involved creative writing tasks such as fan fiction, movie scripts, jokes, or poems.
  2. Homework Assistance (17%+): More than 1 in 6 conversations appeared to be students seeking help with assignments.
  3. Personal Advice (5%): About 5% of conversations involved users asking for personal advice on topics like relationships.
  4. Coding Help (7%): Approximately 7% of conversations requested assistance with writing, debugging, or understanding computer code.
  5. Work-Related Tasks (15%): About 15% of conversations were related to work, including help with presentations and emails.
  6. AI Image Prompts (6%): Around 6% of conversations asked for help creating prompts for AI image generators like Midjourney.
  7. Politeness Factor (13%): About 13% of prompts included the word "please."
  8. User Engagement: While most users interacted with the bots only once, some "superusers" had thousands of conversations over many months.
  9. Coding Specialisation: An additional 1% of conversations were classified as homework help but specifically involved coding assignments.

These statistics offer a glimpse into the diverse ways people are leveraging AI chatbots, from creative and professional tasks to seeking personal advice and academic assistance. As the technology continues to evolve, it will be fascinating to see how these usage patterns shift and what new applications emerge. Read more: Washington Post article


?? This Week's Spotlight: Google NotebookLM

Google NotebookLM is an experimental AI-powered research and note-taking tool developed by Google Labs. Here's what you need to know about this innovative assistant:

What is NotebookLM? It's an AI-powered virtual research assistant designed to help users organise information, gain insights, and develop ideas based on their own sources and notes. Think of it as your personalised AI collaborator that becomes an instant expert in the documents you upload.

Key Features:

  • Source Grounding: NotebookLM allows you to "ground" the AI in specific documents, creating a personalised AI versed in your relevant information.
  • Automatic Summarisation: When you add a document, it generates a summary, key topics, and suggested questions to help you understand the material better.
  • Question Answering: Ask specific questions about your uploaded documents, and NotebookLM provides answers based on the content.
  • Idea Generation: The tool can help you brainstorm new ideas and connections based on your uploaded sources.
  • Citation Support: It provides citations for its responses, showing relevant quotes from the original sources to facilitate fact-checking.

How to Use NotebookLM:

  1. Create a new project and give it a name.
  2. Add sources (currently supports Google Docs, PDFs, Google Slides, and website URLs).
  3. Use the AI to summarise information, ask questions, or generate ideas based on your sources.

Availability and Access:

  • Available to users aged 18 and above in over 180 regions where the Gemini API is accessible.
  • Users can sign up for a free account using their Google login.

Privacy and Data Handling:

  • NotebookLM only has access to the source material that users choose to upload.
  • Files and dialogues with the AI are not visible to other users.
  • Google does not use the data collected to train new AI models.

Limitations and Considerations:

  • It's still an experimental product and may have some bugs or rough edges.
  • Currently limited to 20 sources per project.
  • While source-grounding reduces the risk of AI "hallucinations," users are encouraged to fact-check responses against original sources.

NotebookLM represents Google's effort to reimagine note-taking software with a powerful language model at its core, aiming to help users make connections faster and gain critical insights from their own content.

Ready to give it a try? Visit the Google NotebookLM page to get started!

Read more: Google NotebookLM page | ZDNet | Watch video


?? Generative AI: From Banking to Beyond

I recently came across a fascinating Forrester report about generative AI in banking. While reading it, I couldn't help but think how many of these insights could be super useful for other industries too. So, let's dive into the key takeaways that can benefit businesses across the board!

?? Top Use Cases (Inspired by Banking)

  • Personalised Experiences: AI crafts tailored content and predictive insights, boosting customer loyalty and engagement.
  • Supercharged Customer Service: AI-powered chatbots increase productivity, leading to fewer escalations and happier customers.
  • Smarter Risk Management: AI spots patterns and flags suspicious activity, fighting back against AI-powered fraud (it's like AI inception!).
  • Turbocharged Development: Speeds up coding and documentation, helping modernize legacy systems.


?? Action Steps for Any Industry

⒈ Start Small and Safe:

?? Begin with internal tools to minimise risks

?? Work your way up to customer-facing stuff

⒉ Leverage Existing AI Brainpower

??Use models like GPT-3.5 or GPT-4 as your starting point

?? Teach them your industry's lingo and needs

⒊ Get Your Data House in Order

??Ensure your data can flow smoothly and securely

??Think of it as building a superhighway for your AI

⒋ Set Some Ground Rules

??Create a playbook for responsible AI use

??Always double-check what your AI is saying (trust, but verify!)


?? The Big Picture

Generative AI is like a Swiss Army knife for businesses – super versatile, but you need to learn how to use it right. Start small, get your data ducks in a row, and always keep a human in the loop. The potential is huge, no matter what industry you're in!

Curious to learn more? Check out the full Forrester Report


?? ChatGPT & OpenAI News

  • Advanced Voice Features Showcase: People are pushing the boundaries of what's possible with the ChatGPT App's advanced voice features. Curious to see it in action? Check out these creative experiments: ChatGPT Advanced Voice Mode by Himels Tech | Pushing ChatGPT Advanced Voice to its Limit | on X by Cristiano Giardina
  • Sam Altman Teases 'Project Strawberry' ??: OpenAI's CEO has hinted at a mysterious new AI tool on X (formerly Twitter). While details are scarce, Altman suggests it could be a significant advancement in AI technology. What could it be? Read more: Mashable
  • Structured Outputs for Developers ???: OpenAI has launched a new feature ensuring model outputs perfectly match developer-supplied JSON Schemas. This means: More consistent and reliable AI-generated data, Less time spent on post-processing, Easier integration into existing systems Developers can implement this through OpenAI's Python and Node SDKs. Learn more: OpenAI.
  • GPT-4o System Card Released ??: OpenAI has unveiled detailed information about the capabilities, limitations, and safety measures of their latest GPT-4o model. This transparency is crucial for developers, researchers, and users to understand the model's potential and constraints. Read more: OpenAI | The Verge
  • Rakuten Partners with OpenAI ??: This collaboration aims to enhance customer experiences across Rakuten's platforms. By leveraging OpenAI's models, Rakuten plans to offer more personalized and efficient customer support. Read more: OpenAI | Rekuten
  • AI Safety Expert Joins OpenAI's Board ???: Zico Kolter, a prominent figure in AI safety and machine learning, has joined OpenAI's Board of Directors. His expertise will help guide ethical and secure AI development. Read more: OpenAI
  • Free DALL·E 3 Image Creation: Great news for ChatGPT Free users! You can now create up to two images per day using DALL·E 3 directly within ChatGPT. Just ask ChatGPT to generate an image for you. Read more: OpenAI on Linkedin.
  • MacOS App Update: The latest update introduces a companion window for side-by-side ChatGPT access, enhanced data analysis features, and improved performance. Customize your experience with new settings options. Read more: OpenAI Help.

Stay tuned for more exciting developments from OpenAI!


? Claude Sonnet News

  • AI Safety Bug Bounty Program ???: Anthropic introduces a program encouraging researchers and developers to identify vulnerabilities in its AI models. The goal? Building more secure, reliable, and ethically aligned AI systems. Read more: Anthropic

  • Writing Skills Showdown: ChatGPT vs. Claude ???: A fascinating comparison of these AI giants' writing abilities. ChatGPT shines in versatility and factual accuracy, while Claude impresses with natural, human-like prose. Curious which AI might be the better wordsmith for your needs? Check out the detailed analysis: Geeky Gadgets

Stay tuned for more updates on Claude and Anthropic's innovations in the AI space!


?? Humanoid Robots

  • Figure 02 Ready for Work??: The new Figure 02 humanoid robot is stepping into the workforce, designed to handle tasks in environments where automation is key. With advanced capabilities, it's set to revolutionize industries by taking on jobs that require precision and adaptability. Curious about the future of labor? Read more: The Robot Report Watch more: Figure 02
  • ChatGPT-Powered Humanoid Unveiled??: A new humanoid robot powered by ChatGPT has been introduced, blending advanced AI with robotics. This robot can engage in human-like conversations and perform complex tasks, pushing the boundaries of what AI and robotics can achieve together. Intrigued? Learn more at Geeky Gadgets
  • AI Robot Takes on Table Tennis ??: Google DeepMind's AI-powered robot has entered the world of table tennis! This experiment showcases how far AI has come in mimicking human abilities, with the robot demonstrating impressive speed and precision. Who wins the match? The answer might surprise you. Read more: Digit | Google DeepMind on X
  • Nvidia Controls Humanoid Robots with Apple Vision Pro??: Nvidia has introduced a suite of tools allowing developers to control humanoid robots using Apple Vision Pro. These tools, part of Nvidia's Humanoid Robot Developer Program, include NIM microservices and Nvidia OSMO orchestration service. Developers can now monitor and control robots through gestures and spatial computing, potentially revolutionizing robotic development and operation. Want to know more about this fusion of AR and robotics? Read more: AppleInsider

Excited to see what's next in the world of humanoid robots? Stay tuned for more groundbreaking developments!


?? Corporate AI Developments

  • TikTok's Parent Company Launches AI Text-to-Video Tool ??: ByteDance is shaking things up with a new AI tool that turns text into video. Imagine creating TikTok-worthy content just by typing out your ideas! This could be a game-changer for content creators looking to save time and boost creativity. Ready to see your words come to life? Read more: ZDNet
  • Zoom Unveils AI-Powered Virtual Backgrounds ??: Zoom's stepping up its game with AI-driven virtual backgrounds as part of its new AI Companion features. No more worrying about your messy room during video calls! These smart backgrounds promise to make your meetings more engaging and personalized. Curious about how it'll transform your Zoom experience? Read more: Zoom
  • Yelp's Revenue Rises Amid AI-Driven Ad Sales ??: Yelp's got something to shout about! They've seen a revenue bump, thanks to their AI-powered ad sales, even in a tough market. It's like having a super-smart assistant matching advertisers with their perfect audience. Could this be the future of digital advertising? Read more: PYMNTS
  • Google Scales Back AI in Search Results ??: Google's taking a step back with its AI Overviews in search results. Turns out, sometimes less is more! They're aiming to strike a better balance between AI-generated content and traditional search results. It's all about giving you the most accurate and reliable information. Curious about the impact on your search experience? Read more: Business Insider | What is Google AI Overviews(TechRepublic)
  • AI Business Agent Startup Bardeen Secures Investment from Dropbox and HubSpot ??: Bardeen, the AI whiz kid of business automation, just got a boost from tech giants Dropbox and HubSpot. Their AI agents are all about making your work life easier by handling those pesky repetitive tasks. Could this be the productivity hack we've all been waiting for? Read more: TechCrunch
  • SoundHound Acquires Amelia AI for $80M Amid AI Market Boom ????: SoundHound, the voice AI pioneer, is expanding its repertoire by snagging Amelia AI for a cool $80 million. It's like a tech power couple joining forces! This move could shake things up in industries from banking to healthcare with super-smart conversational AI. Ready for AI that really gets you? Read more: TechCrunch
  • Sequoia Invests in Reflection AI, Founded by Google DeepMind Alumni ??: Sequoia Capital is betting big on Reflection AI, a startup with some serious brainpower from Google DeepMind. Their goal? Creating "superhuman general agents" that could revolutionize knowledge work. It's like having a team of AI geniuses at your fingertips! Excited to see how this could change the way we work? Read more: Business Insider
  • Meet Prompt Poet: Google's New Tool for Revolutionising AI Prompt Engineering ??: Google's latest acquisition, Prompt Poet, is set to make AI prompt engineering a breeze for everyone. It's like having a poet whispering in your ear, helping you craft the perfect AI instructions. Could this be the key to unlocking AI's full potential for both tech wizards and newbies alike? Read more: VentureBeat
  • Apple Could Charge $20 Monthly for 'Apple Intelligence' Features, Analysts Say ??: Apple might be cooking up something special with its "Apple Intelligence" features. Analysts predict a $20 monthly subscription for AI-powered goodies across iOS and macOS. It's like having a super-smart personal assistant for all your Apple devices. Would you be willing to pay for a smarter Apple experience? Read more: CNBC
  • Google Announces Major Updates for Gemini 1.5 Flash and AI Studio ??: Google's giving its Gemini 1.5 Flash model a major boost! We're talking cheaper costs, support for over 100 languages, and cool new features like PDF understanding. It's like they've turbocharged their AI engine for developers. Ready to see what you can build with these upgrades? Read more: Google Developers Blog
  • Google Unveils Gemini-AI Powered TV Streamer for a Smarter Home Experience ??: Google's new TV Streamer is bringing AI right into your living room. Imagine a device that not only streams your favorite shows but also acts as your smart home hub and personal entertainment guru. It's like having a tiny AI wizard in your TV! Excited to make your home entertainment smarter? Read more: The Fast Mode
  • Alibaba Claims Top Spot in AI Math Models with Qwen-2-Math ??: Alibaba's Qwen-2-Math model is flexing its mathematical muscles, outperforming competitors in AI-driven math reasoning. It's like having a math genius in your pocket! Could this be the key to solving complex problems across industries? Read more: VentureBeat
  • You Can Now Ask AI for Your Next YouTube Video Idea ??: YouTube's new AI feature is here to rescue you from creator's block! It's like having a brainstorming buddy who knows all the trends and what your audience loves. Ready to let AI spark your creativity? Read more: TechRadar
  • Palantir and Microsoft Partner to Enhance AI and Analytics for National Security ??: Palantir and Microsoft are joining forces to bring cutting-edge AI and analytics to classified networks. It's like combining a super-spy's toolkit with cloud power! This collaboration could redefine how defense and intelligence agencies use technology. Intrigued by the future of national security tech? Read more: Microsoft News
  • Humane's AI Pin Sees More Returns Than Sales ??: Humane's AI Pin is facing a bumpy start, with more customers sending it back than keeping it. It's a reminder that even the coolest tech needs to win hearts and solve real problems. What do you think it would take for a wearable AI device to succeed? Read more: The Verge
  • Walmart Revolutionises Shopping with AI ??: Walmart's bringing the future of shopping to your fingertips! From virtual try-ons with friends to AI-powered home deliveries, they're transforming how we buy stuff. It's like having a personal shopper and a crystal ball rolled into one. Ready for a smarter shopping experience? Watch it: YouTube - Good Morning America
  • Automattic Launches AI Tool to Improve WordPress Blog Readability ??: Automattic's new AI writing assistant is here to make your WordPress blogs shine! It's like having an editor who never sleeps, helping you craft clearer, more engaging posts. Ready to level up your blogging game? Read more: TechCrunch
  • Amazon Music Introduces AI-Powered 'Topics' Feature for Podcasts ???: Amazon Music's new 'Topics' feature is like having a super-smart podcast guide. It uses AI to help you discover episodes based on specific topics, making podcast exploration a breeze. Ready to dive deeper into your favorite subjects? Read more: Amazon
  • Audible Tests AI-Powered Search Feature with 'Maven' ??: Audible's experimenting with 'Maven', an AI-powered search tool that's like having a librarian who knows your taste in books perfectly. It can even understand natural language queries to find your next great listen. Excited to discover your next favorite audiobook? Read more: TechCrunch
  • Reddit to Test AI-Generated Search Result Pages ??: Reddit's cooking up AI-generated summaries for search results, aiming to help you dive deeper into content and discover new communities. It's like having a friendly guide to the vast world of Reddit. Curious to see how this could change your Reddit experience? Read more: TechCrunch


?? AI Innovations

  • Chinese Team Develops World's First Light-Powered AI Training System ??: A Chinese research team has created a groundbreaking AI training system that operates entirely using light, rather than electricity. This innovative technology could significantly reduce energy consumption and boost the speed of AI training processes, marking a major advancement in photonic computing. Could this be the key to more efficient and powerful AI systems? Read more: South China Morning Post
  • AI Accelerates Drug Development for Mental Health Disorders ??: Researchers at Uppsala University have leveraged AI to predict the 3D structures of key brain receptors, speeding up drug development for mental health disorders. The study focused on the TAAR1 receptor, a promising target for treating schizophrenia and depression. By using AI to model the receptor's structure with high accuracy, scientists can now efficiently identify potential drug molecules. This breakthrough showcases AI's potential to revolutionise traditional drug discovery methods. Ready to see AI tackle mental health challenges? Read more: Phys.org


?? AI in Business

  • Judge Turns to ChatGPT for Case Verdict ????: In a controversial move, a judge used ChatGPT to assist in reaching a case verdict. This decision has ignited a debate on AI's role in the legal system, raising questions about ethics and accuracy. Is AI set to become the next legal advisor? Read more: Daily Mail
  • Verizon Deploys AI to Prevent Internet Outages ??: Verizon is leveraging AI to minimise accidental internet outages through real-time monitoring and analysis of network data. This proactive approach aims to detect and resolve potential issues early, ensuring uninterrupted service for users. Ready for more reliable internet? Read more: ZDNet
  • JPMorgan Chase Develops ChatGPT-Inspired AI Assistant ??: JPMorgan Chase is creating an AI assistant modeled after ChatGPT to help employees access internal research and documents more efficiently. This tool aims to boost productivity by streamlining information retrieval and improving decision-making processes. Could this be the future of workplace efficiency? Read more: CNBC
  • Paige and Microsoft Launch Advanced AI Models for Cancer Diagnosis ??: Paige and Microsoft have unveiled new AI models designed to enhance cancer diagnosis. These models aim to improve both the accuracy and speed of cancer detection, potentially transforming healthcare professionals' approach to diagnosis and treatment. Read more: AI News
  • AI Takes on Baking: Chocolate Chip Cookie Taste Test ??: NPR pitted an AI-generated chocolate chip cookie recipe against a traditional one in a taste test. The AI recipe, created using data from thousands of online recipes, aimed for the perfect flavor balance. While some tasters enjoyed the AI's creation, others still preferred the classic human touch. Would you trust AI with your cookie cravings? Read more: NPR
  • Wendy's Expands AI Drive-Thru Ordering with Spanish Language Option ??: Wendy's is testing a Spanish-language version of its Fresh AI voice agent at 28 locations in Ohio and Florida. The AI system, developed with Google Cloud, has already cut order times by 22 seconds with a 99% accuracy rate. Is this the future of fast-food ordering? Read more: Twitter/X


? Responsible AI

  • Tech Giants Request Extension for EU AI Act Compliance ??: Major tech companies are petitioning the EU for more time to meet the upcoming AI Act compliance deadline. They argue that additional time is needed to fully implement the new regulations aimed at ensuring safe and ethical AI deployment. Will the EU grant this extension, or will tech firms need to scramble to comply? Read more: Cointelegraph
  • Elon Musk's X Halts EU Data Use for AI Training ??: X (formerly Twitter) has agreed to stop using EU user data to train its AI model, Grok. This decision comes amid growing data privacy concerns and increased regulatory scrutiny in Europe. Is this a sign of things to come for other tech giants operating in the EU? Read more: TechCrunch
  • Push for Inclusivity in AI Development ??: AI systems have been shown to replicate societal biases, leading to issues such as gender, racial, and age discrimination. The article emphasizes the importance of female leadership in AI to help mitigate these biases and steer development towards more inclusive technologies. Could diverse leadership be the key to fairer AI? Read more: AI Business
  • Experts Warn of "Addictive Intelligence" Risks ??: As AI systems become more sophisticated and personalised, there's growing concern about their potential to become addictive. This "addictive intelligence" could lead to prolonged, potentially harmful interactions, similar to social media addiction. How can we prepare for and mitigate these risks as AI becomes more integrated into our daily lives? Read more: MIT Technology Review
  • OpenAI Cautions About Emotional Bonds with ChatGPT's Voice Mode ???: OpenAI warns that users might develop emotional attachments to ChatGPT's new voice mode. As AI becomes more lifelike, the boundaries between human and machine interactions blur. Could this reshape our relationships with AI assistants? Read more: Business Standard



??? We'd Love to Hear from You!

Did you find this edition insightful? Like, share your thoughts, and ask us anything.

Stay curious, and until next time, enjoy exploring the vast possibilities AI offers.


???? Subscribe on LinkedIn The ChatGPT Observer Newsletter


As we explore the captivating world of AI technologies, it's important to note that the content of this article is a product of human effort, enhanced by AI tools and additional resources. We have made sure to properly attribute and link all referenced articles to their original sources.


Godwin Josh

Co-Founder of Altrosyn and DIrector at CDTECH | Inventor | Manufacturer

3 个月

The integration of AI into legal systems is poised to revolutionize due process, potentially leading to faster, more data-driven decisions. Imagine a future where AI analyzes case law and precedents with unparalleled speed and accuracy, assisting judges in crafting fairer and more consistent rulings. With the advent of quantum computing, could we envision AI predicting legal outcomes with near-perfect accuracy, raising profound ethical questions about human agency in the justice system?

要查看或添加评论,请登录

Zeyad Sweidan的更多文章

  • The ChatGPT Observer Edition 35

    The ChatGPT Observer Edition 35

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

  • The ChatGPT Observer Edition 34

    The ChatGPT Observer Edition 34

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

  • The ChatGPT Observer Edition 33

    The ChatGPT Observer Edition 33

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

    1 条评论
  • The ChatGPT Observer Edition 32

    The ChatGPT Observer Edition 32

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

  • The ChatGPT Observer Edition 31

    The ChatGPT Observer Edition 31

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story: AI…

  • The ChatGPT Observer Edition 30

    The ChatGPT Observer Edition 30

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

    1 条评论
  • The ChatGPT Observer Edition 29

    The ChatGPT Observer Edition 29

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

  • The ChatGPT Observer Edition 28

    The ChatGPT Observer Edition 28

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: ?? Top Story:…

  • The ChatGPT Observer #27

    The ChatGPT Observer #27

    In a rush? One-minute summary: ?? Top Story: HyperWrite's Reflection 70B, an open-source AI model, sets new benchmarks…

    4 条评论
  • The ChatGPT Observer #26

    The ChatGPT Observer #26

    If you're short on time, here's your one-minute summary of this week's key AI news and advancements: dvancements: ??…

    2 条评论