Happy Wednesday. Tom here! Welcome to Human or Not!

Let’s dive into the latest insights from the world of artificial intelligence.

Here’s what’s on the menu today:

• Meta’s new translator can speak and write in 100 different languages 🗣️

• Can AI-generated music infringe on copyrights? YouTube is working on an answer🎵

• Snapchat wants to create generative AI photos of you 👻

Read time: 3 minutes

TODAY’S PARTNER: Durable

Got a business? Need a website? Don’t have the time? ⏰ 

Build a custom site in just 30 seconds with Durable.

Get your business online in three clicks. Generate a fully designed website with copy, images, and a contact form in under a minute.

Effortlessly customize your site. Simple editing tools, multiple pages, and a free custom domain. No code required.

Start marketing right away. Book more jobs with AI copywriting, SEO, analytics, and a simple AI CRM. All with one login.

🗣️ Meta’s new translator can speak - and write - in 100 different languages

Translation at scale is a notoriously difficult engineering problem, especially across different mediums. One model can translate text to speech, but can’t necessarily translate speech to text.

But Meta’s new AI translation model, SeamlessM4T, can translate 100 languages for speech-to-text and text-to-text, and speech-to-speech. And the model can turn text into speech for 36 different languages.

An incredible amount of data went into this undertaking. Meta mined "tens of billions of sentences" and "four million hours" of speech from publicly available sources to "automatically align more than 443,000 hours of speech with texts, and create about 29,000 hours of speech-to-speech alignments," according to the company’s blog.

This model showcases the power of AI. The manual workload to examine and compile the above data, and make it usable for translation, would have been decades.

Read the full article here:

🎵 Can AI-generated music infringe on copyrights? YouTube has an answer

Can AI-generated music infringe on copyrights? The short answer is yes.

YouTube is embracing AI-generated music, but the company is also being cautious to hedge against potential copyright battles.

YouTube and partners like Universal Music Group (UMG) announced some guiding principles for copyright policies around AI-generated music. But their exact policies aren’t exactly clear yet.

YouTube’s new Music AI Incubator will be the testing ground for AI-generated music, which will help determine the policies.

What we do know is that YouTube will model their new policy after their current system, which allows the rightful copyright owner to make a copyright claim against any video uploaded by another user and receive the monetization proceeds.

In other words, similar copyright rules apply whether content was generated by AI or not.

Read the full article here:

👻 Snapchat wants to create generative AI photos of you

Snapchat is taking selfies to the next level. Their upcoming AI feature, called ‘Dreams’, will create generative AI photos of you and your friends.

Snap can take photos of you and friends and generate new images with ‘fantastical’ imaginary backgrounds and settings.

Now you won’t even need to leave the house to show everyone that you’re doing something cool, like partying at a club or climbing Mount Everest.

But you’ll need your friends’ permission to include them in your imaginary Dreams photos.

Read the full article here:

🗞️ OTHER HEADLINES

• OpenAI brings fine-tuning to GPT-3.5 Turbo

• The AI Trade Faces Its First Major Test With Nvidia Earnings

• Elon Musk's X Follower Count Bloated By Millions Of New, Inactive Accounts

💡TOOLS OF THE DAY

• Scribe: No more “finding time” to document your processes. Scribe automatically creates step-by-step guides for any process. Save hours each week with Scribe.
  

• Clearmind: Turn your complex emotion into clear actions. World's first personalised AI therapy platform to measure and elevate your emotional health.
  

• Aragon: Transform your selfies into AI-generated headshots. Professional photos to make you look your best.
  

• Chapple: All-in-one platform to generate AI content and start making money in minutes. Unlock your business potential by letting the AI work and generate money for you.
  

• Wavel: With real-time AI dubbing technology, you can effortlessly share live streams and meetings with a wider audience at the click of a button.

🎨 AI IMAGE OF THE DAY

Who ordered a Temple?

By @klaudio1993