Darryll Colthrust

Our approach to energy innovation and AI’s environmental footprint An overview of how Google’s approaching AI, energy efficiency and innovation.

Google's Blog Entry: blog.google/outreach-ini...

Google's Technical Report: arxiv.org/abs/2508.15734

I expect this reduction to continue which tends to be the usual practice and pattern with tech. More efficient image generation and editing is likely the next frontier if nano banana is anything to go by. 🍌 I'm looking forward to seeing all other model providers follow suite.

Google have also increased their efficiencies. Over a 12-month period (May 24-May 25) Google achieved significant reductions in their impact of its Gemini Apps:
- 33x reduction in energy consumption per text prompt.
- 44x reduction in carbon footprint per text prompt

However, Google doesn't only serve 1 person. Gemini now has ~400 million monthly active users with typical engagement patterns of about 4 visits per month with around 3 pages viewed per visit. This gives us a rough estimate of over 64 billion queries annually!

⚡ Google has released a comprehensive methodology for measuring the energy, water, and carbon emissions of its AI models in a live production environment across the full infrastructure.

YouTube video by Google Sustainability Calculating our AI energy consumption

⚡ Would you like to know the carbon footprint for Gemini? Read on...

A single median Gemini text prompt:
⚡ Energy Consumption: 0.24 Wh = 9 secs of TV
🌍 Carbon Emissions: 0.03 gCO2e. (e.g. boiling a 1 cup of water is 7g of CO₂e)
💧 Water Consumption: 0.26 mL = 5 drops of water

youtu.be/aarDw3sooYE

💡 I can also see a new set of product features or even new niche sector being created that documents EVERYTHING you do in the creative process to showcase the amount of human input has passed the threshold, whatever that threshold may be in the future.

#AI #Copyright #Technology #Transparency

There's a lot of nuance in this report and in a lot of situations, there is going to be a case-by-case assessment, but human authorship is the current bedrock, but this could be adapted in the future. We'll be rewriting the law on the fly as technology advances.

❌ Prompts alone, even detailed ones, generally do not provide sufficient human control to make users the authors of the output.
❌ Selecting an AI-generated output among uncontrolled options does not equate to authorship.

✅ The inclusion of AI-generated content within a larger human-authored work does not affect the copyrightability of the larger work as a whole.

✅ Human authors can claim copyright if they select, coordinate, and arrange AI-generated material in a creative way. (e.g. Creative selection and arrangement, Creative modifications)

✅ Using AI as a tool to enhance human expression does not limit copyright protection.

✅ When a user inputs their own copyrightable work into an AI system, and that work is perceptible in the output, the user will be considered the author of that portion of the output because it is a derivative.

💡 The US Copyright Office published their 2nd report in the series on AI and Copyrightability. Human+AI can be protected by copyright law.

www.copyright.gov/ai/Copyright...
#AI #Copyright #Technology#Transparency

Turn me into a plushy toy with Google Whisk! You can add other ingredients into the mix for effect. 😁

Similar to Pika 2.0 video ingredients, but this is for images...for now.

www.youtube.com/watch?v=fyFD...

#GoogleWhisk #AI #technology #Google

Currently testing OpenAI's new Sora app. Especially the storyboard feature and it's pretty good! I decided to take some inspiration from the @panmacmillan.bsky.social's Dracula book.

www.panmacmillan.com/authors/bram...

#sora #ai #books #storytelling

🔥Tencent just launched HunyuanVideo, an open-source AI model designed for text-to-video generation with 13 billion parameters. Image-to-video planned.

Project Page: aivideo.hunyuan.tencent.com
Code: github.com/Tencent/Huny...

#AI #video #technology

Is there a link to your paper?