Amazon Science’s Post

"It feels as though we’ve barely absorbed the rapid development and adoption of generative AI technologies such as large language models (LLMs) before the next phenomenon is already upon us, namely agentic AI," writes Warren Center Director Michael Kearns for Amazon Science. The language AI agents might speak, sharing context without compromising privacy, modeling agentic negotiations, and understanding users’ commonsense policies are some of the open scientific questions that researchers in agentic AI will need to grapple with. https://lnkd.in/e4f5GCjq

[News] As AI continues to spread, Morgan State University's Center for Equitable Artificial Intelligence is pushing to address inherent biases in the technology. #AI #bias #tech #highered

Every bubble creates winners and losers. AI will be no different. The difference won’t be who has the fastest tools—it will be who brings the most originality to them. The AI boom feels a lot like past technology bubbles: breathless hype, massive investment, and the promise of transformation. History tells us what comes next—there will be both winners and losers. The winners won’t be the companies that simply plug in the same tools as everyone else. Efficiency gains alone get matched by competitors. What endures is differentiation—the ability to stand out in a sea of sameness. And in an AI-saturated market, that requires human originality: people who know how to use AI not just to accelerate, but to expand the idea space. This is the lesson from every prior wave of disruption: technology levels the field, originality redraws it. The AI bubble will be no different. Hashtags: #AI #OriginalIntelligence #Leadership #Innovation #DigitalTransformation #FutureOfWork

Summary judgment was recently granted in favor of defendants in two copyright cases involving training large language models (LLMs) - Meta’s LLaMA and Anthropic’s Claude. The courts recognized LLM training as a “quintessentially transformative” use. While concerns about AI hurting the market remain, these rulings support continued growth in generative AI and the tech behind it. Read our team's insights: https://lnkd.in/gemY58Xf

The number of evidently AI-generated papers that we are getting nowadays at the Journal of World Intellectual Property is staggering. About 70% of submissions are now flagged as AI, and it goes beyond copy-editing. Hallucinated references, bland articles that say nothing... we're having to work twice as hard to filter out all of the slop.

#Journal of World Intellectual Property is staggering. About 70% of submissions are now flagged as AI, and it goes beyond copy-editing. Hallucinated references, bland articles that say nothing... ---> #Spamming ranked formalization processes, world journals included, to increase probabilities of getting legitimation throw adversarial mishaps.

We asked the PLDF membership for their thoughts on the use of AI in the legal profession. Results published in the newest issue of the PLD Quarterly. https://lnkd.in/g3DeQimN

IPEd has made a formal submission to the Productivity Commission in response to the “Harnessing data and digital technology” public inquiry. As an integral part of the publishing industry and literary sector, we share the deep concerns of writers and publishers about the current and future uses of generative AI, and its potential to undermine the viability and integrity of the creative industries. Read the submission here: https://lnkd.in/gUmKk7pd

Google AI Research and DeepMind have released VaultGemma 1B, the largest open-weight large language model trained entirely with differential privacy (DP). This development is a major step toward building AI models that are both powerful and privacy-preserving.