The Mahazine newsletter #4

Apple & Shutterstock $50 million deal, Google releases imagen 2, IBM and Adobe use AI in LAs Vegas, and Lays highlight sexism in sports with AI

Author

Nicolas De Mahage
April 15, 2024 • Estimated Reading Time: 15 minutes

Welcome to the issue #4 of the Mahazine newsletter

What happened in the AI creative world in April, weeks 1 & 2?

Welcome to the Mahazine, a bimonthly newsletter for people working in the creative industry. In the Mahazine you will find the latest AI news in the creative industry, inspiring visuals and experimentations, innovative ads and marketing campaigns, AI tools and tutorials.

[Note] We moved our newsletter platform from substack to beehiv. If you find this newsletter in your spam folder, please move it to your inbox folder and accept the mail sender to not miss the future issues.

📣 Big news

Apple has reached a major deal worth an estimated $25-50 million to license millions of images from Shutterstock to train its artificial intelligence models, joining other tech giants like Meta, Google, and Amazon who have also struck licensing agreements with Shutterstock recently. This highlights the fierce competition among major tech companies to secure vast troves of data to power their AI systems amid growing privacy concerns around using copyrighted online content without permission for AI training. The burgeoning AI data market, worth billions and projected to grow rapidly, underscores the high stakes as companies vie for an edge in developing cutting-edge AI applications across industries.

Meta will begin labeling a wider range of video, audio and image content as “Made with AI” when detecting an industry standard AI image indicators or when people disclose that they’re uploading AI-generated content.

Google has released Imagen 2, an AI model family, that can generate short video clips from text prompts, aiming to provide a tool for marketers and creatives, though it has limited resolution and capabilities compared to some rival offerings like Runway or Pika. Imagen 2 is available in Vertex AI with inpainting and outpainting features.

Yahoo is acquiring Artifact, the AI-powered news recommendation app created by Instagram co-founders Kevin Systrom and Mike Krieger, in order to integrate Artifact's personalization and content curation technologies into Yahoo News and other Yahoo platforms. While Artifact itself will shut down, its underlying AI tech for categorizing news articles and tailoring recommendations to user interests will be incorporated into Yahoo's products over the coming months, bringing the sophisticated AI features to Yahoo's massive audience of over 185 million monthly news readers.

Stability AI has introduced Stable Audio 2.0, an advanced AI model that can generate high-quality, full-length music tracks up to three minutes long at 44.1kHz stereo from natural language prompts. The new model offers improved capabilities, including audio-to-audio generation, where users can upload and transform audio samples using natural language prompts. The model is trained on a licensed dataset from AudioSparx, ensuring fair compensation for creators and respecting opt-out requests. Stable Audio 2.0 is now available for free on the Stable Audio website and will soon be available on the Stable Audio API.

👀 Creative ads & marketing campaigns

Created by Ogilvy North America, this campaign showcased an animated "Fishy Fish" character made with Adobe's Firefly AI. The cheeky fish appeared everywhere during the Adobe summit - booths, ads, presentations - even in an interactive experience where attendees could "catch" it. The campaign promoted their "Trust What You Create" message around responsible use of generative AI for content.

Lay's conducted a social experiment during a football match in Brazil, using AI technology to change the voice of a female sports journalist to sound like a man. The goal was to highlight the sexism that women face when providing commentary on football. After the match, the journalist revealed the experiment, stating that this is how women are often invalidated, despite their knowledge of the sport. The campaign aimed to promote more respect for women in sports broadcasting and beyond.

Credits: AlmapBBDO

Dove's "Keep Beauty Real" campaign challenges AI unrealistic beauty standards by featuring unedited, diverse images of women. It calls on media and advertising to use more natural, un-retouched imagery. The campaign also provides resources to help build body confidence and resist pressures to conform to narrow beauty ideals. Dove's goal is to shift cultural norms around beauty and empower women to feel confident in their natural selves.

Credits: Soko

Created by le.pub agency, “Playbrary” uses AI to let readers influence the original story, the book becomes playable and generates adventure games of any classic book, from Frankenstein to Sherlock Holmes. Cyril Louis, ECD at Le Pub APAC, said: “Playbrary celebrates books through the lens of technology, inviting a new generation to rediscover the magic of stories in a format that resonates with their digital-first lifestyle.”

Credits: le.pub

🦄 April explorations

Recent visual explorations by us

Fashion editorial shots (Midjourney v6)

🐟🍀🌸

🌞 Curated tweets

BOOST YOUR AI STRATEGY WITH MAHAGE

Elevate your Visual Productions, Digital Projects & Campaigns

Mahage is your strategic partner to leverage the power of AI in your business strategy with :
• AI Strategy & Training
• Creative AI Direction
• AI-Driven Design & Production
• Innovative AI Campaigns

Curious about how AI can skyrocket your business? 👉 We have the answers

🔍 AI Tutorial

Character consistency with Fooocus

Fooocus is an image generating software. Inspired by Stable Diffusion, Midjourney and Leonardo, Fooocus is offline, open source, and free.

Follow this link to install Fooocus

Now, let’s see how it works!

1/ First let’s setup some parameters, you can choose from many styles to get the image type that you want, in our case we will use some Fooocus default parameters and choose Fooocus Photograph style. You can select from Anime, Dark dream, Cinematic and plenty of other styles.

2/ Don’t forget to input a negative prompt if you want to avoid artefacts, wrong styles and effects etc…

3/ Type your prompt and generate an image, in this case: a woman wearing a fur turtle neck sweater. Here is the prompt:

stunning fashion photo profile portrait of a french blond woman, head turned to the left side, neutral background, fur turtle neck sweater

Here is the result photo in higher quality:

4/ Now, you just have to change the prompt slightly, for example change the outfit. Let’s modify fur turtle neck sweater by blue hoodie with red lipstick.

5/ Click on generate, and let the magic happen.

6/ Now it’s your turn to play. You can prompt variations of many parameters to change the makeup, the outfit, the position, the background etc…

Hope you enjoyed this tutorial!

🧠 AI tools

Create and edit full audio song with text prompts using genre, lyrics, and other parameters.

Controllable Video Generation with actual physics understanding

Text-to-image with advanced photorealism and prompt adherence

Transform music into captivating music video

🧑‍💻 Cool new AI features

Spotify 👇 

Spotify has launched AI Playlist in beta, allowing Premium users in the UK and Australia to generate personalized playlists by entering text prompts describing their desired music vibe.

eBay👇

eBay has introduced an AI-powered "shop the look" feature in its iOS app that provides personalized fashion recommendations based on users' shopping history and evolving preferences.

DALL·E👇

The DALL·E editor interface enables you to edit images by selecting an area of the image to edit and describing your changes in chat. You can also provide a prompt with your desired edit in the conversation panel, without using the selection tool.

The editor interface enables you to add, remove, and update parts of your image.

📖 Interesting reads

A snapshot of how 101 of industry leaders are putting AI into production today, creating real-world use cases using Google Cloud AI solutions to enhance their work.

How the metaverse allows architects to design innovative virtual retail spaces free from physical constraints, with brands like Samsung and Balenciaga already creating immersive digital shopping experiences, requiring architects to embrace new skills like storytelling and multi-sensory design.

Created by Knowing Machines, a research project tracing the histories, practices, and politics of how machine learning systems are trained to interpret the world.

📰 Discover the Mahazine monthly issue

Delve deeper into the world of The Mahazine with its monthly digital edition. Enriched with additional visual diaries, creative experiments, and discoveries.

This comprehensive version is exclusively available in a digital format.

If you enjoy the Mahazine, please spread the word by sharing it with your colleagues, friends, and fellow creatives. Your support in growing our community is invaluable as it allows us to reach more creative minds and continue delivering high-quality content.

Reply

or to participate.