🤓 This Week in AI
PLUS: ChatGPT's Hidden Superpower - Effortless File Conversions
Happy Friday!
Did you know you can convert PDFs, Word documents, images, and more directly in ChatGPT? This week, we explore this lesser-known capability that can make tedious file conversions quick and easy for you and your team.
But first it was another big week. My head is still spinning. Here’s what you need to know (clickable links appear in orange in emails and underlined in the Substack app)::
California Governor Gavin Newsom vetoed a bill (SB 1047) to regulate large AI models, citing concerns that it would impose broad, strict safety measures solely based on the size of a model, regardless of risk level. Opponents, including major tech figures, warned the bill could stifle California’s AI industry, a global leader in innovation.
OpenAI has launched Canvas, a new editing tool within ChatGPT that lets users quickly refine specific parts of AI-generated writing or code in a side-by-side view—tweaking only what’s needed without generating entirely new responses.
"Canvas opens automatically when ChatGPT detects a scenario in which it could be helpful. You can also include “use canvas” in your prompt to open canvas and use it to work on an existing project."
Writing shortcuts include:
Suggest edits: ChatGPT offers inline suggestions and feedback.
Adjust the length: Edits the document length to be shorter or longer.
Change reading level: Adjusts the reading level, from Kindergarten to Graduate School, etc.
Add final polish: Checks for grammar, clarity, and consistency.
Add emojis: Adds relevant emojis for emphasis and color.
Here’s what the interface looks like:
Obviously, I had to see what happens when I asked it to write to suit a middle school reading level—because what’s better than keeping things simple? And here’s what it came up with::
I’ll need to play with it lots more but will share my thoughts and insights when I have them.
A direct competitor to Claude’s Artifacts, Canvas rolls out first to Plus and Teams users, with broader access coming soon..
Microsoft's Copilot gets a major upgrade with voice and vision capabilities and a brand-new look and feel. Copilot Voice allows users to chat with the assistant in a more conversational way, while Copilot Vision can interpret text and images on web pages, offering real-time help, like product recommendations or document insights.
It also includes Copilot Daily, an audio summary of news and weather that Copilot reads out as if it were a CNN anchor.
Gemini Live, Google’s answer to ChatGPT’s Advanced Voice Mode, is now available to all Android users, with a planned launch for the iOS app in the future. Live lets you have a back-and-forth conversation with Gemini that features natural exchanges, like being able to interrupt responses with new instructions or information.
So now OpenAI (ChatGPT) has voice AI. Google has voice AI. Meta has voice AI. Microsoft has voice AI. And soon, every startup built on their model will have it too.
So where does this leave us? The end of chat and the keyboard as the dominant way we communicate with computers. Voice will gradually become a more prominent interface with technology.
For businesses, voice AI captures nuanced insights—from tone and context to sentiment—that traditional interfaces miss. This opens the door to hyper-personalized customer experiences and more data-driven decision-making in real-time.
But many companies are not equipped to handle the stronger security risks or the system integration challenges that come with processing this rich data.
The companies that adapt fast will be outcompete in this new voice-first era.
AI workplace tools like Otter.ai, Zoom and Slack—designed to record and transcribe meetings—are unintentionally sharing sensitive conversations, even after meetings end, raising serious privacy concerns.
To avoid these risks:
1️⃣ Carefully review every platform’s privacy policies
2️⃣ Adjust and monitor all AI settings
Google is rolling out ads in AI search summaries to help businesses reach customers at key decision moments with relevant product suggestions. This move gives companies a new way to engage users in real time
But here’s the thing: As a consumer, adding ads to these summaries defeats the whole point. It undermines trust in the information and makes the summaries feel less reliable and more like a sales pitch.
If OpenAI’s upcoming search product remains ad-free as promised, it could position itself as a stronger, more credible and trusted alternative to Google.
People are increasingly turning to AI chatbots like ChatGPT for personal advice—from relationship struggles to medical issues—often revealing sensitive details without realizing it. These conversations are stored by AI companies and could be used for training models or even for targeted ads, raising serious privacy concerns.
Here’s how you can protect your data:
Claude doesn’t use conversations for AI training by default, providing better security.
ChatGPT users need to opt out of data sharing for training. I’ve shared the step-by-step guide to adjust your privacy settings in this newsletter.
Google’s Gemini actively uses user data for training and future products, so I avoid it for any personal or sensitive queries.
ElevenLabs, a leading AI company specializing in hyper-realistic voice cloning is coming for Audible.
Here are some of their latest updates:
✅ Seamless Audiobook Creation: Indie authors can now produce and publish audiobooks directly using ElevenLabs’ AI voice cloning technology. With near-human narration quality, this streamlined process reduces production time while delivering high-quality voice performances—challenging Audible’s dominance and opening the doors to more creators.
✅ Expanding Content Library: The platform’s growing library includes classic books, indie works, and even popular newsletters and blogs. AI-generated celebrity voices bring a unique, immersive experience to listeners, adding a fresh twist to traditional audiobook offerings.
✅ Updated Reader App: Their expanded Reader App—one of my personal favorites— allows users to listen to any article, PDF, ePub, document or text on the go with the highest quality AI voices. You can even paste in ANY content including article links to have it read out loud.
And it seems like they’re just warming up…
ChatGPT maker OpenAI’s latest funding round values the company at $157 billion, reflecting its dominance in AI innovation, but the path ahead isn’t without challenges. Steep development costs, leadership shakeups, and the need to restructure into a for-profit model present critical challenges as the company seeks to sustain its leadership in the AI space.
Nvidia, the world’s top AI chipmaker, has released a powerful open-source AI model—meaning it’s available for anyone to use and build upon. Unlike AI models from companies like OpenAI and Google, which are kept tightly controlled, this move could democratize AI development by giving smaller businesses and researchers access to advanced tools.
Playwright Ayad Akhtar challenges the idea that AI threatens human creativity by using large language models (including ChatGPT, Claude and Gemini) to help write his latest play, McNeal. Currently running at Lincoln Center in New York, the play stars Robert Downey Jr. and explores the allure and risks of AI, questioning whether technology can enhance creativity or redefine it entirely.
AI bots can now pass CAPTCHAs—those puzzles where you identify traffic lights or crosswalks—with 100% success. This makes it much harder for websites to block bots, highlighting a need for new security measures to protect websites from malicious bot activity.
ChatGPT’s Advanced Voice Mode is now available for free accounts but is capped at 15 minutes a month (paid users get about 40 minutes daily). If you are wondering why this mode is a big deal, check out last week’s edition of the newsletter.
It is SOOO good. I’m hooked and I can’t wait until they add vision capabilities and allow document uploads.
AI agents—advanced systems capable of gathering information, learning and taking actions gather information—are changing how organizations make decisions. By automating routine tasks, like researching or scheduling and managing workflows, these agents allow teams to focus on strategic decisions.
PRACTICAL USE-CASES
ChatGPT Makes File Conversions Quick and Easy 🔁
Ever found yourself struggling with file conversions, jumping between different tools, or wasting time looking for a solution?
It’s one of those tasks that’s always tedious and time-consuming. But here’s the good news.
ChatGPT has you covered.
One of its most practical features is the ability to convert files across multiple formats directly within the ChatGPT interface.
Just upload a file, ask ChatGPT to convert it to a new format, and you’re set.
Here are my most common use cases for file conversions:
💠PDFs to Word Docs (and vice versa): Perfect for when you need to edit a text-heavy PDF or finalize a Word document as a PDF for distribution. You can also combine 2 PDFs into one.
💠Image to Text: Handy when you need to extract text from an image, screenshot or scanned document without retyping.
💠Image Conversion (PNG to JPG): Ideal when you need smaller file sizes or better compatibility.
💠ZIP File Uploads: A quick way to upload multiple files or long documents all at once.
The range of supported formats is Impressive. Here are some examples:
📄 Documents: DOC, DOCX, PDF, TXT, HTML (PDF to text only; complex formatting like images, tables, and styles might be lost.)
📊 Spreadsheets: XLS, XLSX, ODS, CSV (conversions don’t retain macros or advanced features)
🎞️ Presentations: PPT, PPTX (basic content only; no animations or embedded media)
🖼️ Images: JPEG, PNG, BMP, TIFF, GIF (format conversion only, no editing capabilities)
📦 Archives: ZIP (up to 100 MB of files) - limited to extracting and compressing only
You can also convert the final draft of whatever you’re working on into a word document, excel spreadsheet, table or a CSV file. Just ask for it in your preferred format and it will provide you with the revised format or a download link.
Tips for Best Results:
Keep it simple with your requests. ChatGPT handles straightforward conversions beautifully but isn't meant for complex formatting.
Double-check the output if your document is heavily formatted. Minor tweaks might be needed post-conversion.
I've saved hours of my time with these simple conversions. Hopefully, you will too.
That's all for this week.
I’ll see you next Friday. Thoughts, feedback and questions are welcome and much appreciated. Shoot me a note at avi@joinsavvyavi.com.
Stay curious,
Avi
💙💙💙 P.S. A huge thank you to my paid subscribers and those of you who share this newsletter with curious friends. It takes me about 8+ hours each week to curate, simplify the complex, and write this newsletter. So, your support means the world to me, as it helps me make this process sustainable.