Data

Google DeepMind, in collaboration with Hugging Face, open sourced its research titled ‘Scalable watermarking for identifying large language model outputs’ in a bid to distinguish between human and AI content on the internet, driven by LLMs. Launched exactly a year ago, SynthID, their watermarking tool, is now available for wider access.

Source: Can Google’s Watermark Tool Save the Internet from AI Slop?

Leading AI companies rely more on content from premium publishers to train their large language models (LLMs) than they publicly admit, according to new research from executives at Ziff Davis. While AI firms generally do not say exactly what data they use for training, executives from Ziff Davis say their analysis of publicly available datasets makes it clear that AI firms rely disproportionately on commercial publishers of news and media websites to train their LLMs.

Source: Ziff Davis study says AI firms rely on publisher data to train models

Meta is reportedly working on AI-powered search to reduce its dependency on Microsoft and Google. The development comes during a season where many vendors are finding new opportunities with search using generative AI. For example, Google — the clear leader in search up to now — has transformed its longstanding search strategy with AI Overviews, a feature that summarizes search results, provides product comparisons and performs other generative AI-powered services, over the past year.

Source: Meta joins others reinventing search with AI | TechTarget

Google’s AI Overviews are expanding across more than 100 countries this week. The AI-generated search summaries will appear for users in Canada, Australia, New Zealand, South Africa, Colombia, Chile, the Phillippines, Nigeria, and many more locations. You can view the full list of countries on Google’s website. In each country, AI Overviews will now appear in any supported language, including English, Hindi, Indonesian, Japanese, Portuguese, and Spanish.

Source: Google’s AI search summaries are rolling out to over 100 more countries

The current crop of generative AI systems needs high quality data, and lots of it. To source this data, big tech companies continually scour the internet, scooping up terabytes of content to feed the machines. But since the advent of widely available and useful generative AI systems in 2022, people are increasingly uploading and sharing content that is made, in part or whole, by AI.

Source: Is generative AI doomed? An expert’s take on the “model collapse” theory