Technology

Researchers from Stanford University have published a report ranking how transparent artificial intelligence (AI) models are and found them wanting. The findings from Stanford Human-Centered Artificial Intelligence (HAI) were compiled into its Foundation Model Transparency Index and gave 10 leading AI models a score out of 100.Among the models that Stanford HAI looked at are Stability AI’s image generator Stable Diffusion, Meta’s Llama 2, and OpenAI’s ChatGPT.

Source: Artificial Intelligence Companies Aren’t Very Transparent, Report Finds

Google has asked a California federal court to dismiss a proposed class action lawsuit that claims the company’s scraping of data to train generative artificial-intelligence systems violates millions of people’s privacy and property rights.  It said the lawsuit would “take a sledgehammer not just to Google’s services but to the very idea of generative AI.”

Source: Google says data-scraping lawsuit would take ‘sledgehammer’ to generative AI

Hollywood studios training AI with preexisting materials could create a whole new set of issues for writers by allowing the studios to use previous work to generate similar materials without the writer’s consent or even awareness. It is in this gray area that thorny issues could sprout, according to Leslie Callif, partner at Beverly Hills entertainment law firm Donaldson & Callif.

Source: Hollywood’s AI issues are far from settled after writers’ labor deal with studios

Artificial intelligence voice cloning software has rapidly increased in quality. It’s allowing anyone from foreign actors to music fans to copy somebody’s voice. On Thursday, a bipartisan group of senators announced a draft bill, called the No Fakes Act, that would penalize people for producing or distributing an AI-generated replica of someone in an audiovisual or voice recording without their consent.

Source: AI voice clones mimic politicians and celebrities, reshaping reality

This feature won’t be locked to just Google Search as the company states you might “see an option to create AI-generated images directly in Google Images”. In that instance, one of the image search results will be replaced with a button offering access to the engine. The creation will slide in from the right in its own sub-window.

Source: Google Search’s generative AI is now able to create images with just a text prompt

Following the wide release of its Firefly generative AI tech last month, Adobe used its annual Max event to introduce new AI-driven features for its creative software, such as Photoshop, Illustrator and Express. Adobe Digital Media CTO Ely Greenfield said the company’s promise of “accountability, responsibility and transparency” where generative AI is concerned “means that we put a lot of effort into identifying, and making sure that we don’t introduce, unintentional harm and bias into the content creation process.”

Source: Adobe Unwraps AI-Driven Tools, Promises “Accountability, Responsibility and Transparency”

At the time of its release, many listeners believed that Ghost’s use of AI to make the song meant that a computer also generated the beat, lyrics or melodies, but as Ghost later explains to me, “It is definitely my songwriting, my production and my voice.” Still, “Heart on My Sleeve” posed pressing ethical questions: For one, how could an artist maintain control over their vocal likeness in this new age of AI?

Source: Ghostwriter, the Mastermind Behind the Viral Drake AI Song, Speaks For the First Time

To prevent the misuse of fraudulent content being presented as authentic, recent developments in generative AI models have made it vital to discern between authentic content and AI-generated material. Watermarking is one method that has been developed to distinguish between photos created by AI and other sources.

Source: Can We Truly Trust Artificial Intelligence AI Watermarking? This AI Paper Unmasks the Vulnerabilities in Current Deepfake Method’s Defense

IMPF’s latest move to ‘guide’ generative music AI comes in the midst of the industry’s rapid acceptance of AI, alongside frequent copyright infringement litigation and discourse surrounding AI models’ lack of transparency. Despite the evident boom of generative AI acceptance and practical use within the music industry, there have been calls to ensure AI developers maintain accurate records of music utilized in training AI.

Source: IMPF Proposes Ethical Guidelines for AI Developers.

In response to the ongoing discourse surrounding the rise of artificial intelligence-generated songs mimicking artists’ voices, the Independent Music Publishers International Forum (IMPF) has issued a new set of guidelines for AI developers. According to a statement, these so-called “ethical principles” around AI, endorsed by 200 independent music publishing companies worldwide, are aimed at ensuring “a more transparent collaboration” between indie publishers and AI tech firms.

Source: Global independent music publishers propose set of ‘ethical’ guidelines for AI developers