• AI KATANA
  • Posts
  • Brands Are Beginning to Turn Against AI

Brands Are Beginning to Turn Against AI

Also: Meta will not launch Meta AI in Europe for now

Author

AI KATANA
June 16, 2024

Morning!

The rapidly evolving landscape of AI continues to generate diverse responses from various sectors. Brands like Dove are distancing themselves from AI in advertising, citing a commitment to authenticity amidst growing criticism of AI-generated content. Meanwhile, digital artist Refik Anadol highlights the creative potential of AI despite facing industry biases. Meta has paused the European launch of its AI assistant due to regulatory concerns, reflecting the broader challenges of data privacy. Pinterest is innovating with “Ad Labs,” exploring AI-enhanced advertising tools, while Instagram’s new AI tagging system has stirred controversy among photographers. In tech advancements, Luma AI’s Dream Machine offers groundbreaking text-to-video generation, and the RoboCasa simulation by UT Austin and NVIDIA advances household robot training. On the venture front, SK Telecom’s $10 million investment in Perplexity AI aims to bolster its AI search capabilities, underscoring the strategic push towards enhanced AI integration. As these developments unfold, they underscore a complex interplay of innovation, ethics, and regulation in the AI domain.

Sliced:

  • ⛔️ Brands Are Beginning to Turn Against AI

  • 🖼️ Refik Anadol on using AI to create a Dior perfume and how digital artists still face bias

  • 🌍 Meta will not launch Meta AI in Europe for now

  • ⚙️ Pinterest has launched Ad Labs, a program to test prototypes of new creative and ad tools

  • 🍎 Apple is handling AI so much better than Microsoft I may ditch Windows for macOS Sequoia

  • ⚠️ Instagram is tagging real photos as 'Made with AI' – and photographers aren't happy

The exuberance surrounding AI-generated imagery is encountering significant backlash from various sectors. Initially celebrated for its potential, AI’s application in media and advertising has faced criticism due to instances of misleading and low-quality content. Brands like Dove have publicly vowed to avoid AI in their campaigns to maintain authenticity and align with their values, while artists and platforms like Cara and PosterSpy are pushing back against the proliferation of AI in creative fields. The gaming and social media industries are also navigating this complex landscape, balancing consumer trust and technological experimentation. As companies grapple with the implications of AI, there is a growing movement towards safeguarding human creativity and transparency, reflecting broader concerns about the ethical and practical impacts of AI.

🖼️ Refik Anadol on using AI to create a Dior perfume and how digital artists still face bias

Digital artist Refik Anadol has been harnessing AI in his work for the past eight years, becoming a pioneering figure in merging art and technology. Known for his large-scale, immersive installations, Anadol’s work blurs the lines between the digital and physical worlds. His use of generative AI, a technique where algorithms create new content based on learned data, allows him to produce vibrant, dynamic visuals that are both awe-inspiring and thought-provoking. Notable projects include a collaboration with Dior to visualize the essence of their J’adore perfume and exhibitions at prestigious venues like MoMA and the Serpentine Galleries. Despite the technological marvels he creates, Anadol acknowledges the biases and challenges digital artists face in a traditionally conservative art world. He views AI not just as a tool for creating new realities but as a mirror reflecting both the possibilities and the complexities of humanity. His optimistic approach contrasts with common dystopian narratives about AI, seeing it as a source of endless creative potential. Anadol’s work encourages institutions to evolve and embrace contemporary digital artistry, advocating for a future where art and AI coexist to enhance human experience.

Meta has postponed the launch of its AI assistant, Meta AI, in Europe following a request from the Irish Data Protection Commission (DPC). This delay is due to concerns raised by European data protection authorities and advocacy groups regarding Meta’s plan to use personal data from Facebook and Instagram to train its AI models without explicit user consent. Despite Meta’s claim that it would only use publicly available and licensed information, the regulatory feedback suggests a need for stricter data handling and privacy measures. Meta expressed disappointment with the decision, arguing that it hampers innovation and competition in AI development within Europe. The company had integrated feedback from European regulators since March, but the recent request from the DPC means that Meta cannot proceed with its plans until it addresses these regulatory concerns. This pause reflects broader challenges and debates around data privacy and AI deployment in Europe.

Pinterest has unveiled “Ad Labs,” a cutting-edge initiative aimed at exploring AI-driven tools for advertisers. The program’s initial focus includes a generative AI model that customizes backgrounds for Product Pins based on users’ aesthetic preferences, creating a more engaging shopping experience. In partnership with brands like Nike and Wayfair, Pinterest is also testing a new feature called Collages, allowing users to curate and shop visually appealing collections that advertisers can promote directly. Early trials of these AI enhancements have demonstrated notable improvements in advertising metrics, such as a 28% reduction in cost per click and significantly higher engagement rates. Additionally, Pinterest is rolling out advanced brand safety measures through collaborations with Integral Ad Science and DoubleVerify, ensuring that ad content remains safe and aligned with brand values. These initiatives reflect Pinterest’s broader strategy to integrate AI seamlessly into its platform, enhancing both user experience and advertising effectiveness while maintaining rigorous standards for brand safety and content quality.

The discussion highlights a shift in how Apple and Microsoft are integrating AI into their operating systems, with a significant focus on user privacy and functionality. Apple’s approach with macOS Sequoia and its AI features, particularly those tied to Siri, emphasizes on-device processing and stringent privacy measures. This strategy ensures user data remains confidential and only resorts to cloud processing when absolutely necessary, and even then, with explicit user consent. In contrast, Microsoft’s aggressive push to embed AI deeply into Windows products like Copilot raises concerns about data security and user trust. This integration often necessitates additional subscriptions to guarantee data protection, which can be off-putting for users wary of their data being shared. The commentary suggests that Apple’s measured and privacy-centric methodology in deploying AI, while initially slower, might offer a more reassuring and user-friendly experience compared to Microsoft’s rapid and more invasive rollout. As a result, even lifelong Windows users are contemplating a switch to macOS due to these growing privacy concerns and the desire for a more controlled AI interaction.

Instagram’s recent implementation of a ‘Made with AI’ tag on photos has stirred controversy among photographers. This tag, intended to indicate AI-generated content, is also being applied to real photos that have undergone minor edits in tools like Adobe Photoshop, which employs AI for some of its features. The move has frustrated photographers who feel that their genuine work is being misrepresented as artificial. Instagram’s Head, Adam Mosseri, explained that the tag is linked to image metadata and reflects any AI involvement, even if it’s just minor editing. This broad application has led to backlash from the creative community, who argue that the label unfairly suggests that the entire photo is artificial. Some suggest a more nuanced tagging system, such as distinguishing between ‘edited with AI’ and ‘created with AI,’ to better inform viewers about the extent of AI use. As the line between real and AI-generated content continues to blur, Instagram’s approach underscores the complexities in managing and communicating the use of AI in digital media.

🛠️ AI tools updates

Luma AI’s Dream Machine is a new text-to-video generation tool that has rapidly garnered attention for its impressive capabilities, offering a competitive edge against established tools like OpenAI’s Sora. Developed by Luma AI, a San Francisco-based startup also known for its 3D model generator Genie, Dream Machine allows users to generate high-quality video content from text prompts for free. This approach democratizes access to advanced AI video creation, inviting widespread experimentation and creativity. Early users have pushed its limits by creating dynamic animations from static images and bringing famous artworks to life with realistic motion. Dream Machine excels in rendering coherent and fluid videos, making it suitable for a wide range of applications from storytelling to music videos and potentially full-length films. The platform’s user-friendly model, combined with plans for a developer-friendly API, positions Dream Machine as a versatile tool for both casual users and professionals. Users can start exploring Dream Machine’s capabilities by signing in through Luma Labs’ website, though current high demand restricts them to ten free video generations per day. As Luma AI continues to refine this technology, it promises to expand the horizons of AI-driven video content creation.

The development of RoboCasa, a comprehensive simulation framework, marks a significant advancement in training general-purpose robots for household chores. Created through collaboration between the University of Texas at Austin and NVIDIA Research, RoboCasa leverages cutting-edge generative AI technologies to create highly realistic and diverse simulated environments. These technologies include text-to-3D models for generating object assets and text-to-image models for enriching environmental textures. The simulation framework offers access to over 2,500 3D objects and 120 detailed scenes, supporting a variety of tasks and interactions with household appliances. This vast array of simulated scenarios allows for the generation of substantial synthetic data, essential for training robots to perform everyday tasks effectively. Despite these advancements, researchers note the ongoing challenge of transferring skills from simulation to real-world applications. Future work aims to enhance the fidelity of these simulations and bridge the gap between virtual training environments and practical, real-world deployment, thereby advancing the capabilities of robots in performing complex, everyday activities autonomously.

💵 Venture Capital updates

SK Telecom has invested $10 million in Perplexity AI, a startup known for its generative AI search engine, marking a strategic move to enhance its AI capabilities. This collaboration aims to improve SK Telecom’s AI personal assistant services, leveraging Perplexity’s technology to fine-tune AI search solutions and integrate Korean data for better localization. The partnership will focus on mitigating the common issue of AI-generated “hallucinations” by incorporating reliable sources such as articles and videos into search results. Perplexity, which processes over 230 million requests monthly and has garnered support from industry giants like Nvidia and Amazon, will also contribute to SK Telecom’s Global AI Platform, fostering mutual growth in AI search services. This initiative was formalized at MWC 2024 in Barcelona and underscores SK Telecom’s commitment to advancing AI applications, offering enhanced, globally competitive search and assistant services.

🫡 Meme of the day

⭐️ Generative AI image of the day