• Everyday AI
  • Posts
  • AI You Can Trust: How reliable data makes it happen

AI You Can Trust: How reliable data makes it happen

AI giants turn to synthetic data, Hyundai and NVIDIA team up, Musk warns of human knowledge cap and more!

Author

Everyday AI
January 09, 2025

Sponsored by

Looking to transform your org with AI? Simply giving employees access won’t cut it.

“That’s like thinking that if we put a treadmill in every home, we’re going to cure heart disease,” says Conor Grennan, Chief AI Architect at the NYU Stern School of Business.

On the latest episode of the WorkLab podcast from Microsoft, he explains the deeper mindset shift AI requires.

It's available wherever you get your podcasts.

Outsmart The Future

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: Your data is your most valuable asset. So how can you harness it safely and effectively with AI? We break it down. Give it a listen.

🕵️‍♂️ Fresh Finds: OpenAI launches newsletter, Google’s new AI podcast feature and xAI releases Grok iOS app. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: AI giants turn to synthetic data, Hyundai and NVIDIA team up and Musk warns that we’ve run out of human knowledge to train AI on. For that and more, read on for Byte Sized News.

🚀 AI In 5: We ranked the top AI tools and features of 2024! See it here

🧠 Learn & Leveraging AI: Pairing AI with your data can be a gold mine IF done correctly and safely. We explain how it’s done. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about Nvidia claims AI chips outpace Moore's Law, U.S. to expand on AI chip restrictions, Meta’s user-generated AI characters gone wrong. Check it here!

 AI You Can Trust - How reliable data makes it happen 📊

Your data is your moat.

Everyone's got AI now.

Find out how reliable data can make your competitive edge happen. Barr Moses, Co-Founder and CEO of Monte Carlo, joins us to discuss.

Join the conversation and ask Jordan and Barr questions on AI and data here.

Also on the pod today:

• Adoption of Data Practices 📊
• GenAI, LLMs, and Data Integration 🔀
• Data Use Case Examples 💼

It’ll be worth your 30 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – Folk is an AI sales assistant and CRM, Sagehood provides AI agents for a 360° analysis of the U.S. stock market and Fernando AI builds apps and websites in minutes.

OpenAI – OpenAI has launched a new newsletter called ChatPT for Education.

Google – Google is testing a new Daily Listen feature that generates a podcast based on your Discover feed.

Cohere AI – Cohere’s co-founder says there’s plenty of opportunity in enterprise and is happy to stay out of the way of ChatGPT.

AI Models - xAI is testing an iOS app for its Grok chatbot.

Microsoft – Microsoft has released a blog on how AI is reducing burnout and transforming frontline efficiency.

AI in Science – An AI model taught on chromatin data can now uncover gene interactions.

1. AI Giants Tap Synthetic Data for Growth 🪴

In a bid to meet the demand for data, Nvidia, Google, and OpenAI are turning to synthetic data factories as traditional data sources dwindle. Nvidia's CEO Jensen Huang highlighted this trend at CES 2025, emphasizing synthetic data's role in advancing AI in automotive and robotics.

Data scientist Ben Lorica notes that synthetic data is becoming pivotal as companies exhaust readily available internet data for training AI models.

2. Hyundai and NVIDIA Team Up for Smart Mobility Innovation 🚗

Hyundai Motor Group revealed a strategic partnership with NVIDIA to advance smart mobility using AI and digital twins. This collaboration aims to revolutionize Hyundai's key mobility products, including autonomous vehicles and robotic systems, by leveraging NVIDIA's powerful computing solutions.

According to Hyundai, the partnership will streamline manufacturing processes, enhance vehicle safety, and improve production quality by utilizing digital twin environments and simulated autonomous driving systems.

3. AI Data Dilemma: Musk Warns of Knowledge Crunch 🧠

In a recent interview on his platform, X, Elon Musk dropped a bombshell, claiming AI companies have maxed out human knowledge for training their models. With traditional data sources running dry, Musk suggests a pivot to synthetic data, despite concerns over AI “hallucinations” muddling accuracy. Andrew Duncan from the Alan Turing Institute echoed these concerns, citing potential risks of "model collapse" and diminishing returns.

As AI pioneers like Meta and Google already dabble in synthetic learning, this shift could reshape how AI is developed, potentially impacting innovation and creativity in the tech industry.

4. NVIDIA CEO Predicts IT Becomes HR with AI Agents 🧑‍💻️

NVIDIA's CEO Jensen Huang made waves at the Consumer Electronics Show by predicting a future where AI agents work alongside humans, transforming IT departments into digital HR teams.

Huang envisions IT professionals managing these AI agents much like HR handles human employees, overseeing tasks from training to ensuring alignment with company culture. As firms integrate AI "digital workers," Chris Daden, CTO at Criteria, forecasts that 30% of companies will harness these agents by 2025 to boost productivity.

5. Perplexity Partners with Tripadvisor for Enhanced Hotel Searches 🏨

Perplexity is integrating with Tripadvisor to offer enriched hotel search experiences, as reported by The Verge. This collaboration means users can now access detailed hotel lists complete with TripAdvisor-sourced summaries, ratings, and images, elevating their search for the perfect stay.

Previously, Perplexity offered basic hotel information, but this upgrade promises a more comprehensive overview with insightful perks like “location” and “service.”

Top AI Tools and Features of 2024

Now that 2024 has come to an end, what were the top AI features and tools?

We decided to rank the biggest and most popular AI tools released in 2024 and give our quick opinion on each feature and tool.

Which ones came out on top?

Find out in today's AI in 5.

In our AI in 5 segment on YouTube, we often create different types of videos. Which are your favorites?

Login or Subscribe to participate in polls.

🦾How You Can Leverage:

When it comes to AI, data can make or break your business. 

Yet, studies show only 33% of business leaders trust their data. 

Record scratch Say whaaaaaa?!

If leaders can’t even trust their own data, then why is the collective biz world rushing to connect LLMs to ….. their data? 

Barr Moses helped us sort the data disaster. 

Barr is the Cofounder and CEO of Monte Carlo, an AI-ready data observability platform. 

She cut it straight today on Everyday AI. 

Working with titans like Fox, Roche, and Credit Karma, Moses sees firsthand how data reliability makes or breaks AI implementation. Her message? 

No data beats bad data. Full stop.

Ready to get your classifications in a row and find out why trustworthy data matters? 

Same. 

Here’s the 3 takeaways you need to know.

1 – The Shift From Model Wars to Data Wars 🚀

Forget about which AI model you're using. That's yesterday's battle. The real fight? It's all about your company’s data you pair up with said model. 

Moses laid it out plain: Every company can access world-class AI models. But only YOU have your company's data. 

That’s the moat.

Here's the truth bomb: A basic AI model with pristine data will smoke a fancy AI model running on garbage data every single time.

Take Credit Karma. Barr said they're not winning because they've got some secret AI sauce. They're winning because they know exactly when you're gonna offer you your next credit card. 

This isn't about data volume. It's about data truth.

Try this:

A KPMG survey found that only 35% of executives fully trust their company’s data. Looks like it's not just your finance and sales teams disagreeing

Create a "data truth" dashboard. Track every single time your data gives conflicting answers about the same thing. You'd be shocked how often finance and sales are living in different realities.

Run a "data lineage" test. Pick your most important metric and trace exactly where that data comes from. If you can't map it end-to-end in 5 minutes, you've got problems.

2 – The Rise of Unrestricted Data Intelligence 💡

Could unstructured data be the secret weapon for AI success?

That's exactly what Barr revealed when discussing her work with insurance companies.

Imagine turning messy customer service chats into structured, actionable data. We're talking about AI that can read between the lines of every customer interaction.

The game isn't just about having clean data anymore. It's about turning chaos into clarity.

One insurance company's doing exactly this – using AI to score customer service interactions on a 10-point scale. 

Genius.

Try this:

Harnessing this data is crucial for AI success.

3 – The New Rules of Data Observability 🔍

Remember when bad data just meant an awkward meeting with your boss? 

Barr said 10-15 years ago, bad data could just be a ‘dang. Got that one wrong. Lolz.” 

Now, bad data can cut Goliaths off at the knees. 

Bad RAG pipelines, as an example, means your AI could be making thousands of wrong decisions per second.

Yikes. 

Barr shared a brilliant example about baseball analytics. A fastball under 80 mph isn't just wrong – it's impossible. But how many impossible things is your data saying right now?

The old way was finding out about data problems from angry customers. The new way? 

Catching that fastball before it even hits the catcher's mitt.

Try this:

Create "impossible data" rules for every key metric. If you're selling shoes, a size 847 probably isn't real. 

Make your system catch that automatically.

Data protection is getting a glow-up with AI. This article shows you how to better think about automated threat detection, governance, and resilience—all in real time.

⌚

Numbers to watch

70%

Microsoft Copilot has been adopted by 70% of Fortune 500 companies since its launch in 2023.

Reply

or to participate.