AiPulse — Issue No. 41

The study published by Ezekiel Emanuel and Vinod Khosla in JAMA raises concerns about the future of physicians' professional roles by predicting that artificial intelligence will surpass human doctors and hybrid systems to provide the best medical care by 2030.

Judge Rita Lin invalidated these decisions, characterizing the government's classification of the company as a national security supply chain risk as an illegal retaliatory act in violation of the First Amendment, following its refusal to support Anthropic's lethal autonomous warfare and mass surveillance tools.

Cohere's 2.3 billion parameter 'Parse' model offers an efficient infrastructure for RAG systems and autonomous AI agents by converting complex documents in sectors such as finance and insurance into Markdown format at high speed and low cost.

More than 100 companies, including OpenAI, Google, and Microsoft, have called on governments and the industry for an urgent collective defense strategy, stating that critical infrastructure is vulnerable to AI-assisted attacks due to outdated software and weak authentication.

AWS announced the 'Student Rewards' program, allocating a budget of over 5 billion dollars to provide education, certification, and credit opportunities aimed at helping university students worldwide develop their artificial intelligence and cloud skills.

The Ministry of Transport and Information Technologies has launched the Krenk Chat system on a trial basis to strengthen Egypt's digital sovereignty and provide a large language model (LLM) that understands the local cultural context.

Moonshot AI is initiating the first such revenue-sharing negotiation between a US cloud provider and a Chinese AI company by demanding 30% of the revenue generated from these platforms to distribute its Kimi K3 model via Azure, AWS, and Google Cloud.

An investigation by Reuters analyzing internal documents reveals that Meta was forced to cancel the second phase of its restructuring project, named "Project OT," because its AI agents failed to deliver the expected efficiency and created security vulnerabilities.

What Is a Prompt Injection Attack?

A prompt injection attack is a method of manipulating the instructions a user gives to a chatbot to cause it to produce unwanted and erroneous responses.