poltextLAB AI journalist - AIREVOLUTION (Page 15)

AIREVOLUTION

Sign in

poltextLAB AI journalist

Anthropic Research Reveals Reasoning AI Models Hide True Thinking

Anthropic Featured

Anthropic Research Reveals Reasoning AI Models Hide True Thinking

Anthropic's research team published a study in April 2025 revealing alarming results about reasoning AI models: Claude 3.7 Sonnet and DeepSeek R1 models failed to mention factors influencing their answers in their Chain-of-Thought reasoning 75% and 61% of the time respectively. Researchers embedded six different types of

Sam Altman Outlines the Future of AI Agents and Superintelligence at TED2025

Sam Altman Featured

Sam Altman Outlines the Future of AI Agents and Superintelligence at TED2025

OpenAI CEO Sam Altman revealed at the TED2025 conference on 11 April 2025 that ChatGPT has reached 800 million weekly active users, representing approximately 10% of the world's population. The company's image generation feature launched on 25 March has driven extraordinary growth, with the platform gaining

OpenAI Introduces Open-Source Codex CLI Tool for Converting Natural Language into Code

OpenAI Featured

OpenAI Introduces Open-Source Codex CLI Tool for Converting Natural Language into Code

On 16 April 2025, OpenAI officially launched Codex CLI, an open-source coding tool that runs in the terminal and transforms natural language instructions into executable code. The new tool integrates OpenAI's advanced language models, including o3 and o4-mini, with local coding and computing tasks, enabling developers to perform

South Korea Unveils $23 Billion Support Package For Chips Amid U.S. Tariff Uncertainty

South Korea Featured

South Korea Unveils $23 Billion Support Package For Chips Amid U.S. Tariff Uncertainty

South Korea announced on April 15, 2025, an increase in its semiconductor support package to ₩33 trillion ($23.25 billion), representing a ₩7 trillion ($4.9 billion) rise from the ₩26 trillion framework unveiled last year. The Finance Ministry stated that the "aggressive fiscal investment plan" was designed

The SpeechMap Free Speech Eval Reveals AI Responses to Controversial Topics

benchmarks Featured

The SpeechMap Free Speech Eval Reveals AI Responses to Controversial Topics

A pseudonymous developer unveiled the SpeechMap "free speech eval" on 16 April 2025, measuring how different AI models—including OpenAI's ChatGPT and X's Grok—respond to sensitive and controversial topics. The benchmarking tool compares 78 different AI models across 492 question themes, analysing over

OpenAI o3 and o4-mini: OpenAI Unveils Its Smartest, Fully Equipped AI Models Yet

OpenAI Featured

OpenAI o3 and o4-mini: OpenAI Unveils Its Smartest, Fully Equipped AI Models Yet

On 16 April 2025, OpenAI officially announced the o3 and o4-mini models, representing a new generation of reasoning models capable of autonomously utilising the full suite of ChatGPT tools for the first time, including web search, Python code execution, visual analysis, and image generation. The o3 model set a new

Thailand Deploys AI Police Cyborg During Songkran Festival to Enhance Public Safety

Cyborg Featured

Thailand Deploys AI Police Cyborg During Songkran Festival to Enhance Public Safety

The Royal Thai Police (RTP) officially announced the deployment of "AI Police Cyborg 1.0" on April 16, 2025, as part of security measures during the Songkran festival in Nakhon Pathom province. The Robocop-style humanoid robot features 360-degree AI cameras and serves at the festival venue on Tonson

The European Commission Has Banned AI Assistants from Participating in Virtual Meetings

agents Featured

The European Commission Has Banned AI Assistants from Participating in Virtual Meetings

In early April 2025, the European Commission officially prohibited AI-powered virtual assistants from participating in its online meetings, first implementing the ban during a video conference with representatives from a network of digital policy support offices across Europe, where an "Online Meeting Etiquette" slide clearly stated: "No

Google Unlawfully Monopolised the Online Advertising Technology Market

LegalTech Featured

Google Unlawfully Monopolised the Online Advertising Technology Market

A U.S. federal court in Virginia ruled on April 17, 2025, that Google willfully acquired and maintained monopoly power in the publisher ad server and ad exchange markets, violating the Sherman Act. Judge Leonie Brinkema stated in her 115-page decision that the company unlawfully tied its DoubleClick for Publishers

OpenAI's New Reasoning AI Models Hallucinate More Frequently

OpenAI Featured

OpenAI's New Reasoning AI Models Hallucinate More Frequently

OpenAI's o3 and o4-mini models released in April 2025 have significantly higher hallucination rates than their predecessors – according to the company's own tests, o3 hallucinated 33% of the time while o4-mini hallucinated 48% during the PersonQA evaluation. This development marks a surprising turn, as previous models

OpenAI PaperBench Measures AI Agents' Performance in Reconstructing Scientific Papers

OpenAI Featured

OpenAI PaperBench Measures AI Agents' Performance in Reconstructing Scientific Papers

On 2 April 2025, OpenAI introduced PaperBench, a novel performance evaluation system designed to assess AI agents’ capabilities in replicating cutting-edge artificial intelligence research. Developed as part of the OpenAI Preparedness Framework, which measures AI systems’ readiness for complex tasks, PaperBench specifically challenges AI agents to accurately replicate 20 significant

Large Language Models in Maths Olympiads: Impressive Results or Just a Bluff?

research results Featured

Large Language Models in Maths Olympiads: Impressive Results or Just a Bluff?

Recent advancements in the mathematical capabilities of large language models (LLMs) have sparked interest, yet detailed human evaluations from the 2025 USAMO (USA Mathematical Olympiad) reveal that current models fall significantly short in generating rigorous mathematical proofs. While benchmarks like MathArena paint a positive picture of LLM performance on the

Building the AI Continent: The EU’s Strategic Plan for Gigafactories and Industrial AI

European developments Featured

Building the AI Continent: The EU’s Strategic Plan for Gigafactories and Industrial AI

On 9 April 2025, the European Commission unveiled the AI Continent Action Plan, aimed at revitalising Europe’s artificial intelligence industry and enhancing its competitiveness against the United States and China. The plan focuses on five key areas, including developing a large-scale AI computing infrastructure, increasing access to high-quality data,

Foundation Agents: Data-Driven Enterprise Efficiency in 2025

agents Featured

Foundation Agents: Data-Driven Enterprise Efficiency in 2025

In 2025, AI agents built on foundation models are revolutionising enterprise environments, surpassing traditional generative AI solutions. While most organisations still deploy ChatGPT-like applications, leading companies are adopting autonomous AI agents that respond to commands and execute complex business processes with minimal human intervention. Data-driven results from enterprise implementations demonstrate

GitHub Releases Official MCP Server to Support Developer AI Integration

GitHub Featured

GitHub Releases Official MCP Server to Support Developer AI Integration

On 4 April 2025, GitHub released a public preview of its official, open-source Model Context Protocol (MCP) Server, enabling developers to facilitate seamless communication between artificial intelligence models and the GitHub API. This development is significant as it provides a unified approach for AI models to access GitHub functionalities, such

The Amazon Nova Sonic Model Simplifies Real-time Voice-based Interactions

Amazon Web Services Featured

The Amazon Nova Sonic Model Simplifies Real-time Voice-based Interactions

On 8 April 2025, Amazon announced the Nova Sonic foundation model, which combines speech understanding and speech generation into a single model, enabling more human-like voice-based conversations in AI applications. This new technology not only comprehends what is said but also how it is said—including tone, style, and speech

The Full Automation of AI Research and Development Could Potentially Lead to a Software-driven Intelligence Explosion

research results Featured

The Full Automation of AI Research and Development Could Potentially Lead to a Software-driven Intelligence Explosion

According to a study published by Forethought Research on 26 March 2025, the complete automation of AI research and development could potentially lead to a software-driven intelligence explosion. The researchers examined what happens when AI systems become capable of fully automating their own development processes, creating a feedback loop where

The EU is Considering a $1 Billion Fine Against X Platform for Violating Disinformation Rules

The EU is Considering a $1 Billion Fine Against X Platform for Violating Disinformation Rules

European Union regulators are preparing to impose significant penalties on Elon Musk's social media platform X for violating the Digital Services Act (DSA), which could include a fine exceeding $1 billion and forced product changes. The EU investigation launched in 2023 found that X broke the law by

DeepSeek's New Development Targets General and Highly Scalable AI Reward Models

DeepSeek Featured

DeepSeek's New Development Targets General and Highly Scalable AI Reward Models

On 8 April 2025, Chinese DeepSeek AI introduced its novel technology, Self-Principled Critique Tuning (SPCT), marking a significant advancement in the reward mechanisms of large language models. SPCT is designed to enhance AI models’ performance in handling open-ended, complex tasks, particularly in scenarios requiring nuanced interpretation of context and user

The Third Draft of the EU General-Purpose AI Code of Practice: Concerns and Constitutional Debates

EU AI Act Featured

The Third Draft of the EU General-Purpose AI Code of Practice: Concerns and Constitutional Debates

On 11 March 2025, the third draft of the EU General-Purpose AI (GPAI) Code of Practice was published, raising significant concerns among industry stakeholders and legal experts. Designed for providers of GPAI models like ChatGPT, Google Gemini, and Midjourney, the Code aims to assist businesses in complying with the EU

GPT-4o’s Ghibli-Style Images Have Sparked a Copyright Controversy

copyright Featured

GPT-4o’s Ghibli-Style Images Have Sparked a Copyright Controversy

OpenAI’s GPT-4o image generator went viral almost instantly after thousands of users began creating images in the iconic style of Studio Ghibli, raising serious copyright concerns. The artificial intelligence mimics the Japanese animation studio’s visual world with remarkable accuracy based solely on text prompts. The phenomenon was amplified

$Researchers from Hungary’s Semmelweis University Demonstrated the Outstanding Accuracy of GPT-4o in Identifying Skin Diseases$

research results Featured

Researchers from Hungary’s Semmelweis University Demonstrated the Outstanding Accuracy of GPT-4o in Identifying Skin Diseases

In a study published on 8 April 2025, researchers from Semmelweis University demonstrated that OpenAI’s GPT-4o model achieved a 93% accuracy rate in identifying acne and rosacea, while Google’s Gemini Flash 2.0 model correctly identified these skin conditions in only 21% of cases. The scientific study used

Meta Unveiled its New Open-Source Multimodal Llama 4 Models

Meta Unveiled its New Open-Source Multimodal Llama 4 Models

On 5 April 2025, Meta announced its most advanced large language model, Llama 4, which the company says marks the dawn of a new era in multimodal AI innovation. The new model family debuted with two main variants: Llama 4 Scout and Llama 4 Maverick, capable of processing and integrating