Trending AI news stories + paper

Today's Top AI news stories and papers

AI news stories

1. Ex-Google engineer charged with stealing AI secrets
   Linwei Ding, a former Google engineer, faces charges for allegedly stealing AI-related trade secrets while working for two Chinese tech firms. Accused of transferring over 500 files and taking a CTO role in China without Google's knowledge, Ding risks up to 10 years in prison per charge. The case highlights tensions between the U.S. and China over technology and trade, with U.S. officials emphasizing national security and innovation protection.

2. Salesforce announces new AI tools for doctors

   Salesforce has introduced new AI tools to reduce administrative workload for healthcare professionals, notably with Einstein Copilot: Health Actions for scheduling, patient data summarization, referral processes, and Assessment Generation for digital health assessment creation. These innovations are part of the Einstein 1 Platform, aiming to centralize diverse medical data sources and alleviate physician burnout by streamlining cumbersome clerical tasks. The tools are set to be HIPAA-compliant and will be rolled out progressively, with full availability by the end of the year.

3. Google experiments with a tool to enable on-device AI

   Google has introduced the MediaPipe LLM Inference API, enabling developers to run AI models directly on devices such as laptops and phones despite their limited computing capabilities compared to servers. This innovation allows large language models to operate fully on-device, a significant advancement given their substantial memory and computational requirements. The MediaPipe framework supports various models and is compatible with multiple platforms, with plans for further expansion.

4. Microsoft engineer warns company’s AI tool creates violent, sexual images, ignores copyrights

   Shane Jones, a Microsoft engineer, has raised concerns about the company's AI image generator, Copilot Designer, creating violent and sexual content, as well as potentially breaching copyright laws. Despite his efforts to alert the company and propose that the product be taken off the market or that its usage be restricted to mature audiences, his suggestions have not been adopted. Jones escalated the issue by writing to the FTC and Microsoft's board, highlighting the lack of adequate safeguards and the tool's potential to generate harmful content.

5. Amazon’s generative AI bot Rufus makes online shopping easier (for the most part)

   Amazon's new generative AI chatbot, Rufus, is designed to enhance the shopping experience on its app by providing detailed information and answering questions about products, thereby making shopping more efficient and less reliant on external searches. Although Rufus generally aids in making informed purchasing decisions, it can sometimes become confused, as demonstrated when differentiating between product types. Despite these limitations, Rufus represents a significant step in integrating AI into consumer shopping, offering potential benefits such as increased engagement within the Amazon app and improved customer decision-making.

6. AI2 Incubator scores $200M in compute to feed needy AI startups

   The AI2 Incubator, initially part of the Allen Institute for AI and now an independent entity, has received $200 million in computing resources to support AI startups, enabling them to accelerate their development significantly. This substantial support will help these startups by providing them with up to $1 million in dedicated AI compute resources, addressing their critical need for computing power to develop and scale their innovative solutions. Managed by Jacob Colker, the AI2 Incubator focuses on aiding pre-seed startups, leveraging this significant compute allocation to help them rapidly progress toward revenue generation and advance their technologies.

AI Repos

1. VAST-AI-Research / TripoSR

2. hpcaitech / Open-Sora

3. alibaba-damo-academy / FunASR