AI News

Pink Floyd's recent choice of a fan-made AI video for their "Dark Side of the Moon" anniversary sparked controversy in the art community. Critics lamented the undermining of traditional artistry, while the winning entry, created by Damian Gaume, showcased AI's impact on creativity. This debate highlights ongoing challenges artists face regarding AI's role in their work.

Intel and Altera have launched AI-optimized processors and FPGAs for edge computing to boost industries like retail, healthcare, and automotive. The new Intel Core Ultra and Atom processors, alongside Agilex 5 FPGAs, enhance AI integration for improved on-premises decision-making and efficiency.

Stability AI has launched Stable LM 2, a 12-billion-parameter language model, enhancing its generative AI offerings beyond its renowned image technology. This model, designed for improved performance and multilingual conversation, aims to balance power with usability. It follows updates to its smaller counterpart and emphasizes safety and compatibility with other tools.

Apple's licensing agreement with Shutterstock allows access to millions of images for AI training, valued between $25 million and $50 million. This move positions Apple alongside other tech giants in the competitive AI landscape, amid growing privacy and copyright concerns over data usage.

The U.S. Department of Commerce will provide TSMC $6.6 billion for chip manufacturing facilities in Arizona, part of the CHIPS and Science Act. This investment, combined with Intel's and GlobalFoundries' contributions, aims to enhance U.S. semiconductor production, create jobs, and secure supply chains. TSMC's initiatives will significantly bolster the U.S. position in the global semiconductor market.

Spotify has launched AI Playlist, a beta feature for Premium users that allows personalized playlist creation through prompts, now available in the UK and Australia. Users can input various elements, preview playlists, and save them. This follows the rollout of the AI DJ feature.

Data scientists and AI professionals face significant risks if AI systems are not secured during development, including dataset poisoning and prompt injection attacks. Collaboration with cybersecurity specialists, routine audits, and adherence to regulatory guidelines are essential to mitigate vulnerabilities and protect sensitive data.

Meta is expanding its labeling of AI-generated content to include a "Made with AI" badge for media that meets certain AI indicators or is voluntarily disclosed by users. This change responds to Oversight Board recommendations, aiming for transparency in manipulated media ahead of the 2024 elections, while maintaining community standards.

Cohere launched its enterprise AI model, Command R+, claiming it outperforms OpenAI's GPT-4 in benchmarks. Available on Microsoft Azure and Amazon SageMaker, pricing varies. Microsoft promotes Azure's extensive model library as it competes with Amazon's offerings, while Cohere aims to expand its customer base.

The open-source Ray framework, utilized by many organizations including for training GPT-3, was compromised by the “ShadowRay” vulnerability (CVE-2023-48022), exposing sensitive data for seven months. Anyscale has since released tools to check for open ports, highlighting the need for secure practices as infrastructure attacks increase.

SiMa.ai has secured $70 million in funding, with significant backing from Dell Technologies Capital, highlighting the growing interest in edge AI technology. SiMa.ai's integrated MLSoC technology aligns with Dell's edge solutions, promising advancements in sectors like retail, logistics, and healthcare. As enterprises seek to leverage AI at the edge, this investment positions SiMa.ai for potential growth and innovation in the edge computing market.

Gretel has launched the largest open-source Text-to-SQL dataset, featuring over 100,000 synthetic samples to enhance AI model training. This initiative addresses data access challenges across industries while prioritizing data privacy and quality, empowering businesses with improved AI capabilities.

Introducing Genspark: The Next Generation of AI-Powered Search

The Media AI Impact Tour in San Francisco, led by Matt Marshall, explored generative AI's applications in business. Key industry leaders discussed moving from initial experiments to critical implementations, addressing relevancy, cost, and the importance of retrieval-augmented generation (RAG) for accurate data integration.

A class-action lawsuit by artists against AI companies over copyright claims has gained momentum, with major defendants like Stability AI and Midjourney seeking dismissal. They argue that their models create new works and lack evidence of copyright infringement. The case’s evolution follows a previous dismissal of many artists’ claims.