OpenAI joins Meta in the fight against deepfake images 💥

Welcome to another edition of Horizon AI,

With deepfakes and AI-powered misinformation grabbing headlines, companies are eager to demonstrate their commitment to addressing the issue. Meta took the lead yesterday by announcing measures to label AI-generated images on its platforms, and just hours later, OpenAI is following suit with a similar initiative.

Let's jump into it!

Read Time: 3.0 min

Here's what's new today in the Horizon AI

• OpenAI is Adding New Watermarks to DALL-E 3 Images 💥

• AI Research: Apple releases 'MGIE', an AI model for instruction-based image editing

• AI Tutorial: Create animated videos in minutes using AI

• AI Tools to check out

• AI Image of The Day🎨: Reader-Submitted Images

• The Latest in AI and Tech 💡

AI News

OPENAI

OpenAI is Adding New Watermarks to DALL-E 3 Images 💥

OpenAI has just announced that it will start applying metadata labels to any imagery created by its AI tools. This will allow any user to identify the images as AI-generated.

Details:

• The move comes just hours after Meta announced a similar measure to label AI-generated images across its platforms. Highlighting the growing industry-wide concern for AI-powered misinformation

• The metadata tagging change is effective immediately for web users of ChatGPT and DALL-E 3 and will be implemented for all mobile users by February 12.

• Accompanying the update, OpenAI has provided a link to Content Credentials, a platform where users can upload images to check whether they are AI-generated.

It's important to note that this change applies only to newly generated images with ChatGPT and DALL-E 3. Images created before this update will not feature the new metadata

Source

AI RESEARCH

Apple’s 'MGIE', an AI model for instruction-based image editing

Apple has released an exciting new AI tool called MGIE (MLLM-Guided Image Editing) that allows users to edit photos through natural language instructions. This technology aims to simplify image editing using the power of large language models.

Details:

• Leverages capabilities of multimodal large language models (MLLMs) to interpret text commands and make pixel-level edits with remarkable accuracy. Rather than needing to use complicated editing tools, users can just describe what they want to do to the image, and MGIE will automatically make those changes.

• The model offers a diverse array of editing capabilities, including Photoshop-style modification, global photo optimization, and local editing.

• MGIE is accessible as an open-source project on GitHub. Users can also try it out through a web demo hosted on Hugging Face Spaces.

This new form of AI-enabled image editing has the potential to make image editing much easier and more accessible for everyone.

Source

AI Tutorial

Create animated videos in minutes using AI

Thanks to AI, you can transform your ideas into dynamic characters.

The tool we’ll be using is Viggle, which allows you to create animated videos in multiple ways:

• Create a character using prompts and make it move

• Animate your existing character with motion templates

• Combine your character with a reference video

To use this tool:

1. Join Viggle’s discord and select a “Creation” channel

2. Use /character to create a character

3. Click the motion button below the image and select the kind of animation you want and you will get your video featuring your character in action.

4. If you already have a character that you want to animate, type /animate, upload your character image, and choose a motion_template

5. Or type /mix if you want to make a video from scratch. You will have to upload the character and the motion you want it to imitate

Source: @shushant_l on X

AI Tools to check out

Agenthub: Automate any workflow with AI, no code required.

Grammarly: AI writing assistant. It reviews spelling, grammar, punctuation, clarity, engagement, and more.

Meetext: AI-powered note-taking tool that effortlessly summarizes classes, meetings, and talks for enhanced efficiency.

Supadash: Connect your database and instantly get AI-generated charts to visualize your data and track metrics

Leafo: Use in-app painting or take photo of your drawings and bring it alive with more than 15 motions and custom backgrounds.

AI Image of The Day

Send your AI-generated images for a chance to get published

The latest in AI and Tech

Huawei bets on AI over bestselling phones

Faced with production constraints, Huawei is prioritizing its Ascend AI chips over its popular Mate 60 phones. The chips power AI model training, an area of increasing domestic demand. By focusing on the Ascend 910B chip, Huawei aims to meet this growing appetite. The move indicates Huawei’s belief in AI’s future growth, even if it means slowing the output of the Mate 60, which beat iPhone sales in China last year.

Read more

Lawyer faces $8K fine for fake ChatGPT Case

A New York lawyer who cited a fictional case generated by ChatGPT in a malpractice lawsuit not only got her client's case dismissed but now faces an $8,000 penalty. She admitted to consulting the AI tool when unable to find a real case to support her claims. But as ChatGPT's tendency to "hallucinate" detailed but false information shows, it's not reliable for legal research.

Read more

Britain makes 100 million pounds investment in AI research and regulation

The UK announced substantial funding, totaling over 100 million pounds, for AI research and regulation initiatives. Nearly 90 million pounds would go towards nine specialized research hubs exploring applications in healthcare, chemistry, mathematics, and ethical AI practices in partnership with the United States. 10 million pounds are aimed at equipping UK regulators with practical tools and frameworks for monitoring AI risks across crucial sectors.

Read more

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻