• AI PlanetX
  • Posts
  • Google’s Reimagine Tool Creates Disturbing Images

Google’s Reimagine Tool Creates Disturbing Images

D-ID Unveils AI Video Translator with Voice Cloning

Author

Altiam Kabir
August 23, 2024

Welcome to another edition of AI PlanetX

Google’s Reimagine tool produces unsettling images; D-ID rolls out AI video translator with voice cloning.

In Today's Edition:💎

  • Hottest AI News

  • Top AI & SaaS Tools

  • AI Tutorial

  • Top AI & Tech News

  • Interesting Uses of AI

  • Top AI Video Tutorial

  • AI Job List

  • Free AI Course of the Day

Hottest AI News

Google

Pixel 9 Series' New Feature Excels at Creating Unsettling Images, Raising Concerns

AI photo editing takes a concerning turn with Google's new "Reimagine" feature on the Pixel 9 series. This tool lets users blur reality and fiction in images, raising serious ethical questions about digital imagery's future.

Details:

  • The feature can generate highly realistic additions to photos, including disturbing elements like car wrecks, corpses, and drug paraphernalia.

  • While safeguards exist, they can be circumvented with careful wording, allowing users to create potentially harmful or misleading content.

  • Edited images lack obvious indicators of manipulation, with only easily removable metadata tags to identify AI-generated content.

This development underscores the urgent need for robust detection tools and heightened digital literacy to navigate an increasingly complex visual landscape.

D-ID

New AI Video Translator by D-ID Features Voice Cloning and Lip Sync

AI is revolutionizing video translation, with D-ID's new tool translating videos, cloning voices, and syncing lip movements, transforming content creation for global audiences.

Details:

  • Translates videos into 30 languages, preserving voice and syncing lip movements.

  • Uses the Rosetta-1 model for AI Video Translate, handling videos from 10 seconds to 5 minutes and up to 2GB.

  • Builds on D-ID's photo animation success, backed by $25M in Series B funding for U.S. enterprises.

D-ID’s Video Translation tech is accessible through D-ID Studio and API, with a one-month free trial available. It works best with one person facing the camera.

Top AI & SaaS Tools

  • Periodix (Lifetime Deal): AI autopilot manages all your LinkedIn outreach, from crafting messages to scheduling calls

  • Hexus: AI creator for engaging product demos & docs

  • Phi-3.5: The new models include Phi-3.5-mini-instruct (3.82B) for basic tasks, Phi-3.5-MoE-instruct (41.9B) for advanced tasks, and Phi-3.5-vision-instruct (4.15B) for image/video analysis

  • Ideogram 2.0 (NEW): Text-to-image model with new styles, color palette control, and exceptional text rendering [FREE]

  • Midjourney: The web app is now open to all with free trials available again

AI Tutorial

How to Fine-Tune GPT-4o

OpenAI launched fine-tuning for GPT-4o. Until September 23, they offer 1M free training tokens daily for GPT-4o and 2M daily for GPT-4o mini.

Follow these steps to get started:

  • Navigate to the OpenAI Fine-Tuning Console and click on Create.

  • Choose "gpt-4o-2024-08-06" from the Base Model options.

  • Submit your training dataset—essentially, a collection of sample conversations.

Remember, you can't upload raw text or PDFs for ChatGPT to reference.

Your data must be in JSON format for ChatGPT to understand it. Though it may sound challenging, formatting is simple. Each example should include:

  • A "system" message (defining GPT's role)

  • A "user" message (the input or question)

  • An "assistant" message (GPT's expected response)

Here's an example of how it should be structured:

{"messages": [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "What's the largest planet in our solar system?"}, {"role": "assistant", "content": "The largest planet in our solar system is Jupiter."}]}

Ensure that each example is on a separate line and follows the correct JSON format.

Provide at least 10 examples, but 50-100 is recommended for best results.

Once done, visit the OpenAI Playground to interact with your customized model and evaluate its performance.

Top AI & Tech News

  • Workers at Google DeepMind push company to drop military contracts [Read More]

  • Microsoft is rebranding Copilot in the most Microsoft way possible [Read More]

  • OpenAI joins opposition to California AI safety bill [Read More]

  • McAfee rolls out deepfake detector in Lenovo's new Copilot-Plus PCs [Read More]

  • Perplexity AI plans to start running ads in fourth quarter [Read More]

  • Microsoft’s revised Recall AI feature will roll out to beta testers in October [Read More]

Interesting Uses of AI

AI Art Spotlight

AI Model: Ideogram 2.0

Prompt:

An illustration of a white t-shirt with bold gold and red text that says "what a glorious time to be free". The text is written in a vintage, cursive font. The background is a light blue sky with fluffy white clouds.

AI Prompt of the Day

Content Marketing Ideas

This prompt is about suggesting [number] content marketing ideas that will create valuable content and drive engagement for the agency’s client.

Prompt:

Suggest [number] content marketing ideas that can create valuable content and drive engagement for our agency's client.

Context:

Client industry — [client industry here]
Target audience — [client's target audience here]
Current content marketing efforts — [describe current content marketing efforts here]

Inspiration:

"[content marketing idea one]"
"[content marketing idea two]"
"[content marketing idea three]"

Content guidelines:

"[your content guidelines here]"

Top AI Video Tutorial

Create AI VFX with Runway Gen-3

AI Job List

Free AI Course of the Day

Microsoft Azure AI Fundamentals: Generative AI

In this course, you'll learn about language models in generative AI, discover Azure AI Studio's capabilities, see how copilots boost efficiency, and understand Microsoft's approach to ethical AI.