AI-Fakes: How to Spot Them

Welcome to another edition of the best damn newsletter in AI.

As a reminder, we’ve opened up Reader Questions. Our mission is to make using AI accessible to every business, and a huge part of that is making sure you’re not stuck and that we continue answering diverse questions. So, submit your question here. We’ll review every single one and cover answers here in the newsletter.

Here's what we're covering today:

• When verification is better than detection

• 2 hacks to help you feel better about AI-fakes

• Summarize YouTube videos with AI

Let’s get to it! 👇

TODAY’S PERSPECTIVE

Detection < Verification: 2 Hacks to Feel Better about AI-Fakes

This week Katy Perry fooled everyone with an AI-generated image of her at the Met Gala… including her mom!

If you feel like we are barreling towards a world that feels like Black Mirror, you’re not alone. 

“You shouldn’t believe everything you read on the internet”.

Is now… 

“You shouldn’t believe everything you read… any image you see, any video you watch, any voice you hear… on the internet, on TV, over the radio, in a song, texting you, emailing you, or even calling you”. 

*Gulp. 😅 

Yes, big tech companies like OpenAI, Adobe and Meta are working on AI detection methods via watermarking, using AI to detect AI and other projects.

But with detection, you’re reliant on a 3rd parties’ ability to detect and tell you that something is AI-generated.

Today we want to help you shift your mindset. Take the matter into your own hands. 

Verification is usually a better strategy than a Detection.

Any time you are predicting something, you should assess the cost of being wrong.

If the cost of a false positive is higher than a false negative → you can rely on detection. But if the cost of a false negative is higher than a false positive → you’re better off adopting a strategy of verification.

So how do you verify?

2 hacks to help you verify what you see online:

Skill #1: Google Reverse Image Search (quickest!)

Did you know you can see all of the places that an image is used on the internet? It takes 3 seconds, and can help you quickly understand the internet’s footprint of an image.

Right click on an image and Select “Search Image with Google”. In Chrome, you’ll see a side panel pop up with other instances of that image online and videos, and you can easily check the other sources.

Skill #2: Learn the SIFT Method (taught in leading universities & used by journalists!)

If you’re looking for something more robust than a reverse image search, and that will work across any type of information, look no further than the SIFT Method. When applied, these 4 moves will make you confident that you’re looking at fact not fiction:

S: STOP

I: Investigate the Source

F: Find Better Coverage

T: Trace Claims to their Original Context

Here’s a free course we liked on how to apply SIFT to verify information. 

USE CASE

Effortlessly Summarize YouTube Videos with Video Summarizer GPT

Ever wished you could get the gist of a lengthy YouTube video without having to watch the whole thing? We've discovered a CustomGPT that can summarize any YouTube video with just a link.

Here’s how it works:

1. Go to the Video Summarizer GPT in ChatGPT.

2. Enter the link of the video you want summarized.

3. Ask it to summarize the video for you.

4. You can even chat with the video by asking questions about specific content you want more information about.

This feature is free for ChatGPT Plus subscribers. Now, you can save time and get straight to the point!

LINKS

For your reading list 📚

AI is bringing new opportunities to the business world…

• Anthropic launches a new premium plan aimed at businesses to compete with OpenAI's ChatGPT Enterprise.

• X launches Stories, delivering news summarized by Grok AI as a clever way to summarize trending stories without partnerships.

• Microsoft bans U.S. police departments from using enterprise AI tool in a policy change after Axon announced plans to use GPT-4 to summarize audio from police body cameras.

AI is making strides in scientific research and environmental protection...

• CMU is developing low-flying drones to map wildfires as a new approach to provide better information against increasingly severe wildfires.

• Sperm whale 'alphabet' discovered, thanks to machine learning as researchers from MIT CSAIL and Project CETI used machine learning to analyze a large dataset of sperm whale vocalizations.

AI is reshaping cultural trends and artistic expressions...

• Teens are making friends with AI chatbots.

• Randy Travis gets his voice back in a new Warner AI music experiment using AI software and vocals by singer James DuPre.

Have a burning AI question? Or stuck trying to help your team adopt AI?

We read every single question submitted and use these to inform the topics we go deep on. 👇️ Submit your question below!

That's all!

We'll see you again soon! Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected].

Cheers,

🪄 The AI Exchange Team