✋YouTube slams OpenAI for stealing data for Sora

JOB BOARD | PRO | PODCAST | TOOLS | ACCREDITATIONS

Welcome to AI Tool Report!

Friday’s top story: YouTube CEO has blasted OpenAI for potentially stealing YouTube data to train its text-to-video model Sora.

🎙️ Head over to our podcast channel to get a quick round-up of the top news stories in AI this week.

🌤️ This Morning on AI Tool Report

✋YouTube slams OpenAI for stealing data for Sora
⛔ Google, IBM, and Microsoft rally to stop AI job losses
🪴 OpenAI is expanding its custom model program
🪧 New York defends chatbot giving illegal advice

Read Time: 5 minutes

FINANCE

DATA

✋YouTube slams OpenAI for stealing data for Sora

Our Report: Despite a lack of concrete evidence that OpenAI is using YouTube data to train Sora, its text-to-video model, YouTube CEO—Neal Mohan—has declared that it will be a “clear violation” of its terms of service if it has.

🔑 Key Points:

Mohan states that when a creator uploads their work to YouTube, they expect their terms of service to be met, and having their content used by a third party is “a violation” of this.
OpenAI CTO (Mira Murati) said Sora was trained on "publicly available data and licensed data," but worryingly didn't know if that included videos from YouTube.
This comes as OpenAI has previously admitted that it has used copyrighted data to train its models because it’s “impossible” to build the technology without it.

🤨 Why you should care: While there are clear ethical issues with this approach, Google is developing its own AI tools, with Mohan assuring users that “Google adheres to individual contracts with creators before using any YouTube videos”, meaning it’s using YouTube data to train its own models, but is refusing to let its competitors do the same.

In Partnership with OctoML

Watch: Learn to Futureproof your GenAI Stack

Running GenAI at scale is hard! And there’s a frenzy of choice between new models and how to run them. In the face of rapid change, we all want the same thing: efficient, customizable, and reliable inference to power our apps.

Watch this lively discussion between AI insiders, visionary investor Allie K Miller, and OctoAI CEO Luis Ceze to learn how to build your stack to maximize new open-source models with your data in an environment you control.

What they cover:

Evaluating and managing multiple models
Fine-tuning and model asset management
Smarter hardware utilization and lowering TCO
Data privacy and security

Go watch this talk then sign up for a virtual technical deep-dive event covering GenAI in your private environment.

JOB LOSS

⛔ Google, IBM, and Microsoft rally to stop AI job losses

Some of the biggest global tech companies—including Google, IBM, and Microsoft—have formed a consortium to address the impact of AI-related job losses.
The consortium (AI-Enabled Information and Communication Technology Workforce Consortium) will upskill or reskill employees who have lost (or could lose) their jobs.
This comes as estimates show that 30% of back-office jobs will be automated by AI, causing 25% of employees to expect AI-related layoffs, with 4,000 employees already replaced.

Prompt Engineering

Use the Prompt Engineer GPT by AI Tool report to 10x your prompts.

✅ Copywriting

^{copy & paste ⬇️}

“I need a [type of text] that will persuade [ideal customer persona] to purchase my [product/service] by highlighting its unique benefits and addressing any potential objections.”

ENTERPRISE

🪴 OpenAI is expanding its custom model program

OpenAI is expanding its Custom Model program, which gives Enterprises access to OpenAI researchers who can help train AI models for specific use cases and applications.
With dozens enrolling, OpenAI now plans to introduce “assisted fine-tuning” and “custom-trained models” to the program to further bolster model performance on specific tasks.
This extended program saw Law firm, Harvey, create a custom AI-powered tool that incorporated billions of legal texts and attorney feedback to speed up case preparation.

CHATBOTS

🪧 New York defends chatbot giving illegal advice

New York City Mayor (Eric Adams) has defended ‘MyCity’—the first city-wide chatbot to give business owners “trusted information”—even after it gave wrong and illegal advice.
It wrongly advised employers that they could take their workers’ tips, and claimed that they didn’t need to give their employees notice if their schedule changed.
MyCity is still live, with Adams proclaiming that it was a pilot program—launched in October—so “you need to put it into the real environment to iron out the kinks.”