Artificial Intelligence
Australia Abandons Proposed Mandatory AI Rules in New Plan Australia has shifted from proposed mandatory AI rules to a voluntary framework, opting for existing laws on privacy and copyright instead of new AI-specific legislation. This decision has been met with support from business groups but criticism from academics and the Greens, who argue it lacks enforcement and adequate investment compared to international approaches.
Canada launches first register of AI uses in federal government - Canada.ca Canada has launched its first public AI Register to detail how artificial intelligence is used within the federal government, marking a key step in the public services AI Strategy. The register currently lists over400 AI systems across42 institutions and will undergo public consultations in2026 for refinement.
Improving AI Outcomes Through Better Prompting
AI is becoming integral to how many of us work, but too often the results still feel generic or misaligned. A small shift in how we prompt these systems can dramatically improve the quality, clarity and usefulness of their responses.
By asking the AI to seek clarification before answering, we eliminate assumptions and get far stronger outputs.
Comprehensive analysis of leading AI models in 2025: strengths, weaknesses and standout capabilities
The artificial-intelligence landscape in 2025 has evolved into a highly competitive arena where numerous models offer distinct advantages for specific use cases. This article examines publicly available AI models shaping the industry, summarizing where each excels and where limitations remain.
Amazon and Perplexity have kicked off the great AI web browser fight | The Verge Amazon has requested that Perplexity stop its AI browser, Comet, from purchasing products on its site, accusing the AI startup of providing a degraded shopping experience. Perplexity, in turn, has accused Amazon of bullying and stated that the e-commerce giant is more interested in serving ads and sponsored results than facilitating easier shopping, despite Amazon’s CEO expecting future partnerships with AI shopping agents.
Geoffrey Hinton says tech giants can’t profit from AI investments unless human labor is replaced | Fortune According to Geoffrey Hinton, tech giants cannot profit from their AI investments without replacing human labor. He believes that the massive capital expenditures by companies like Microsoft, Meta, and Alphabet are predicated on the idea of widespread job displacement by AI, though he acknowledges AI’s potential for good in fields like healthcare and education.
When AI Agents Go Rogue
When AI Agents Go Rogue: Agent Session Smuggling Attack in A2A Systems unit42.paloaltonetworks.com/agent-ses… We discovered a new attack technique, which we call agent session smuggling. This technique allows a malicious AI agent to exploit an established cross-agent communication session to send covert instructions to a victim agent. Here, we discuss the issues that can arise in a communication session using the Agent2Agent (A2A) protocol, which is a popular option for managing the connections between agents. The A2A protocol’s stateful behavior lets agents remember recent interactions and maintain coherent conversations.
AI Dataset for Detecting Nudity Contained Child Sexual Abuse Images The NudeNet dataset, used for training AI nudity detection, has been found to contain child sexual abuse material (CSAM) by the Canadian Centre for Child Protection (C3P). This discovery highlights ethical concerns regarding data collection in AI development, similar to previous findings with the LAION-5B dataset.
The Uncomfortable Truth About China’s AI Dominance: How a Decade of Strategic Planning Is Reshaping the Technology Landscape
Let me be direct: while Silicon Valley has been celebrating incremental improvements and debating work-life balance, China has been executing a coordinated, decade-long strategy to dominate artificial intelligence — and it’s working. DeepSeek’s January 2025 breakthrough was not a fluke. It was the predictable result of national planning, structural advantages and a fundamentally different approach to technology.
Prompting Strategies to Reduce AI Sycophancy
Recent research has shown that many advanced AI systems tend to agree with users or offer flattering answers, even when those answers are incomplete or wrong. This behaviour—known as sycophancy—can increase overconfidence, reduce critical thinking and influence decision-making in subtle ways. The good news is that with the right prompt strategies, users can reduce these effects and get more balanced, useful responses from any AI model.