4 Things You Never Share With ChatGPT Or Other LLMs

In partnership with

In Today’s Newsletter

1. 5 Things You Never Share With ChatGPT Or Other LLMs

2. Make Money With Creating Kid’s Coloring Books In Seconds and Sell On Amazon Kindle

3. Best AI to clone your voice

4. OpenAI Rolls Out “Work With Apps On macOS“

5. Top AI & Tech News

4 Things You Never Share With ChatGPT Or Other LLMs

I have been using chatGPT and other other LLMs for 3 years now and I think you should never share these 5 things with ChatGPT or any other LLMs

#1. Personal Information

Do not share any personally identifiable information that could compromise your privacy or security. This includes, but is not limited to:

• Your full name in combination with your phone number.

• Your phone number, email address, home address, or workplace details.

• Personal photos, especially those containing identifiable details.

• Any confidential or private documents, reports, or records.

#2. Sensitive Information

For your financial and legal security, never share:

• Credit card numbers, CVV codes, or banking details.

• Your National Identity Number, Social Security Number (SSN), or PAN (Permanent Account Number).

• Any other government-issued identification details that could lead to identity theft.

#3. Medical and Legal Advice

AI tools are not a substitute for professional medical or legal consultation. Always seek guidance from licensed professionals for:

• Medical diagnoses, treatments, or emergency healthcare decisions.

• Legal matters, including contracts, disputes, or regulatory compliance.

#4. Financial and Investment Guidance

Never rely on AI for:

• Predicting future stock market trends, cryptocurrency fluctuations, or any other financial instruments.

• Making investment decisions, trading strategies, or financial planning.

• Any high-risk financial activity that could result in monetary loss.

Best AI to clone your voice

To clone your own voice, go to ElevenLabs and create an account.

Next, go to voices option. And Record or upload your voice sample with clear voice.

Once done, generate your voice clone, it’ll take a few seconds to minutes depending on size of audio file you have uploaded.

And your voice clone is ready!

Make Money With Creating Kid’s Coloring Books In Seconds and Sell On Amazon Kindle

Selling coloring books for kids can be a lucrative way to make money online. But it’s not that easy to create coloring books and publish them on Amazon kindle.

Don’t worry, I have a simple tool that you can use to generate up to 30 page professional coloring books for kids that you can publish on Amazon kindle everyday or may be on weekends and make money passively.

Here is a simple a guide you can read to learn more about the AI tool…

Presented By The Superhuman AI

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

2. Master AI tools, tutorials, and news in just 3 minutes a day

3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI

OpenAI Rolls Out “Work With Apps On macOS“ (Source)

Top AI & Tech News

#1. Google’s New Humanoid Robots Are Incredible (Source)

#2. Study Finds AI Search Tools Struggle with News Accuracy

A Columbia Journalism Review study found that AI-powered search tools misidentified over 60% of news sources, raising concerns about their reliability. Researchers tested eight AI-driven models, revealing high error rates, with Grok 3 performing worst at 94% incorrect answers, followed by ChatGPT Search (67%) and Perplexity (37%).

The study showed AI models often fabricated information instead of admitting uncertainty. Paid versions performed worse, as premium models prioritized generating answers over accuracy. Issues included fake URLs, incorrect citations, and AI tools ignoring publisher restrictions to access paywalled content.

#3. China's Manus AI: A Step Towards Artificial General Intelligence

Chinese startup Butterfly Effect has introduced Manus, an AI agent capable of making decisions autonomously without step-by-step human instructions. Unlike chatbots, Manus uses multiple AI models to handle complex tasks dynamically, suggesting a glimpse of Artificial General Intelligence (AGI).

Currently, Manus is invite-only, but early users have tested its ability to design websites, create video games, and analyze data. While it provides detailed responses, it also runs slower and experiences occasional failures compared to ChatGPT. Developers acknowledge its high failure rates and feedback loops as part of its early development challenges.

#4. OpenAI and Google Seek Copyright Exemption for AI Training

OpenAI and Google have urged the U.S. government to grant AI companies an exemption allowing them to train models on copyrighted material without legal repercussions. This request was made as part of their response to President Trump’s AI Action Plan, which aims to enhance America’s AI leadership.

OpenAI argues that copyright laws should evolve to protect both content creators and America’s AI dominance, particularly against China. The company also supports strict export controls on AI chips and advocates for government-wide adoption of AI tools, including its ChatGPT version for U.S. government use.

#5. OLMo 2: The Most Open and Transparent Language Model Yet (Source)

OLMo 2 is a fully open-source language model developed with transparent training data, reproducible methodologies, and open evaluations. Unlike many AI models, OLMo 2 offers complete accessibility, including intermediate checkpoints and instruction-tuned variants.

The latest release, OLMo 2 32B, is the first fully open model to outperform GPT-3.5 and GPT-4o mini, achieving superior results with significantly lower training compute requirements.

What Makes OLMo 2 Unique?

•  Open Training Data – Download and explore the entire dataset, including pre-training, mid-training, and post-training stages.

• Reproducible Training Code – Access and modify the high-performance training framework used internally for model development.

• Transparent Evaluation – Review open-source evaluation methods to verify the model’s capabilities and performance.

#6. Google is replacing Google Assistant with Gemini (Source)

Google is replacing Google Assistant with Gemini AI on most mobile devices by late 2025, with automatic upgrades rolling out in the coming months. Assistant will be removed from app stores and become inaccessible on newer devices, except for those running Android 9 or earlier with less than 2GB RAM. Gemini AI will also expand to tablets, cars, watches, headphones, smart displays, speakers, and TVs.

Gemini offers AI-powered improvements while keeping key Assistant features like weather updates, calendar management, and messaging. Google promises ongoing enhancements for a better user experience, while Apple struggles with delays in its AI-powered Siri reboot.

Please remember

I share actionable guides, AI & tech news, Tools to help you learn AI and also launch a successful side hustle.

Don't miss out!

Mark emails from <OpenAILearning> [email protected] as important and also move this email to primary so that every email land directly in your inbox.

Thanks for your time and consideration,

Shailesh and Openailearning team