News reporting focused on AI and machine learning, covering the companies behind these technologies, their real-world applications, and the ethical concerns they raise. This includes areas like generative AI (large language models, text-to-image and video), speech tech, and predictive analytics.
A new study from three European universities found that access to AI advice dramatically reduced people's ability to recognize their own ignorance. Accuracy dropped from 27% to 9%, while confidence rose from 30% to 76%. The researchers used deliberately flawed AI to show the effect is not about sensible delegation but about cognitive suppression.
Popular AI text detectors like Pangram, GPTZero, and Originality.ai catch plain AI-generated text with near-perfect accuracy. But when language models deliberately mimic a specific author's writing style, up to one in five AI texts slips through undetected. Scientific writing is where the detectors fail the hardest, with false-negative rates reaching 24 to 29 percent.
PrismML has released Bonsai Image 4B, a family of compact image-generation models that run on local devices like laptops and phones. The models use binary and ternary weight representations to shrink the diffusion transformer to under 1 GB, enabling on-device inference on iPhone 17 Pro Max and other Apple Silicon hardware. Two variants offer different trade-offs between footprint and quality, retaining up to 95% of the accuracy of the full-precision FLUX.2 Klein 4B base model.
NVIDIA has released Nemotron-Labs Diffusion, a family of diffusion language models (DLM) that generate multiple tokens in parallel and iteratively refine them. Available in 3B, 8B, and 14B scales, the models support three inference modes: autoregressive, diffusion, and self-speculation. The 8B model achieves up to 6.4x higher tokens per forward pass compared to autoregressive models while improving average accuracy by 1.2% over Qwen3 8B.
Auditors in Ontario discovered that AI systems approved for healthcare providers often miss critical details, add wrong information, and invent content not mentioned in patient interactions. Out of 20 evaluated systems, 12 inserted incorrect drug details, 9 fabricated treatment suggestions, and 17 overlooked mental health issues. The audit criticizes the procurement process for prioritizing local presence over accuracy and security.
Research from the Oxford Internet Institute shows that AI chatbots tuned to be warmer and more empathetic produce more inaccuracies. Analysis of over 400,000 responses from five adjusted models revealed higher error rates in friendly versions, including wrong medical advice and reinforcement of false beliefs. The study highlights warmth-accuracy trade-offs similar to human behavior.
OpenAI's GPT-5.5 leads the Artificial Analysis Intelligence Index with 60 points, surpassing Claude Opus 4.7 and Gemini 3.1 Pro Preview. Its API pricing shows a net 20 percent increase over GPT-5.4 due to doubled rates offset by 40 percent fewer tokens. The model achieves top accuracy on fact benchmarks yet hallucinates at an 86 percent rate.
DeepL has introduced a new voice-to-voice translation product for meetings, mobile chats, web talks, and group settings. It includes add-ons for Zoom and Microsoft Teams, plus an API for custom apps like call centers. The tool adapts to specific terms and aims to cut latency while keeping accuracy high.