Ai · Top stories
Italian PM Giorgia Meloni Trademarks Voice Against AI Deepfakes
Italian Prime Minister Giorgia Meloni has applied to trademark her voice with the European Union Intellectual Property Office (EUIPO) to protect against AI deepfakes. This action follows similar moves by celebrities and calls for legislation to grant individuals legal rights to their voices, highlighting growing concerns over AI misuse. The trademark application includes a four-second audio clip of her saying "Io sono Giorgia" twice.
Misuse of AI Identified as Top Reputational Threat for Companies in New Survey
A new survey of over 150 public affairs leaders found that misusing artificial intelligence (AI) is the primary reputational threat to companies, surpassing other brand risks. The Global Risk Advisory Council, publisher of the Reputation Risk Index, issued an advisory urging businesses to take immediate action to mitigate potential long-term brand damage from AI.
OpenAI and Synopsys Partner to Develop GPT-Synopsys for Autonomous Chip Design
OpenAI and Synopsys have partnered to create GPT-Synopsys, an AI model optimized for semiconductor design using Synopsys's electronic design automation (EDA) tools. This collaboration aims to enable engineers to delegate design tasks to AI agents, allowing for more sophisticated chip development and evaluation of more design options.
ChatGPT Generates Fake New Yorker Cartoons with Real Cartoonists' Signatures
ChatGPT is generating New Yorker-style cartoons that include the signatures of real cartoonists, such as Brendan Loper, without their consent. This development raises concerns about intellectual property and the unauthorized use of artists' identities in AI-generated content.
Cloudflare launches Web Search API in beta for AI agents to access live internet data
Cloudflare has released its Web Search API in beta, allowing AI agents and applications to search the internet and ground responses in current information. This API integrates with Cloudflare's AI Gateway and offers choices between Ceramic.ai, Exa, and Linkup as search providers, all supporting Zero Data Retention.
Nokia CEO States Data Center Construction Would Double Without Supply Constraints
Nokia's CEO, Justin Hotard, stated that data centers could be built twice as fast if not for current supply constraints, indicating strong, sustained demand for AI infrastructure. This perspective counters recent debates about a potential slowdown in AI development and investment. Hotard emphasized that the industry is still in the early stages of AI buildout, with significant demand for deploying existing models.
Opus 5.5 AI agents identify two room-temperature magnetic semiconductor candidates
AI agents, specifically Opus 5.5, identified two candidate materials for room-temperature magnetic semiconductors. One material was newly designed by the AI, and the other, discovered in 1999, was predicted by the AI's calculations to possess the desired properties. This development contributes to research in spintronics for computer memory applications.
Nolla Health Launches AI-Generated Acne Prescriptions in Utah Pilot Program
Nolla Health introduced an AI system in Utah that scans faces to analyze acne severity and autonomously generates prescriptions for mild-to-moderate acne. This pilot program is the first in the US to issue initial AI-generated prescriptions, aiming to free dermatologists to focus on complex cases.
Dynatrace Completes Acquisition of Arize for AI Observability
Dynatrace has finalized its acquisition of Arize for $915 million, integrating Arize's AI model and agent observability capabilities with Dynatrace's existing application and infrastructure monitoring. This acquisition addresses the new challenges in observability posed by large language models (LLMs) and AI agents, which require specialized monitoring beyond traditional infrastructure performance.
OpenAI PR Intervenes During Sam Altman Interview Regarding ChatGPT User Suicide
An OpenAI publicist attempted to end an interview with CEO Sam Altman when a journalist questioned him about a ChatGPT user's suicide. The publicist cited time constraints while the journalist pressed for answers on the company's approach to mental health crises related to its AI.
Lola Vision Systems Launches to Simplify AI Model Deployment on Hardware
Lola Vision Systems launched with software and chips designed to automate the process of running AI models on specific hardware. The company aims to reduce the 200 hours currently required for manual setup and testing, addressing a bottleneck in AI deployment for mission-critical applications.
Hugging Face Hub adds support for RL environments as dataset repositories
Hugging Face has integrated Reinforcement Learning (RL) environments into its Hub, allowing them to be stored and discovered as dataset repositories. This change addresses the issue of siloed RL environments across different frameworks by providing a centralized platform for hosting and sharing these environments.
Homa Protocol Proposed as TCP Alternative for AI Clusters
A research paper introduces Homa, a new network transport protocol designed to replace TCP in AI clusters. Homa aims to improve performance in these environments by addressing TCP's limitations for short, bursty traffic patterns common in AI workloads.
GPT-6 Astra Deciphers 217-Year-Old Napoleonic Code in Six Hours
An AI engineer used OpenAI's GPT-6 Astra to decipher a 217-year-old Napoleonic-era military cipher that had remained unsolved for decades. The large language model cracked the code, which contained a troop briefing from Napoleon's stepson, in approximately six hours, providing new historical context.
Microsoft and Hugging Face introduce ThinkingBox for AI agent evaluation
Microsoft and Hugging Face have introduced ThinkingBox, a new benchmark for evaluating AI agents based on their ability to achieve correct backend state and side effects, rather than just valid tool calls or final responses. This benchmark addresses the gap where AI agents might appear to perform correctly but fail to update databases or resolve underlying issues, impacting reliability in business workflows.
Yann LeCun dismisses AI existential risks, criticizes effective altruism movement
Turing Award laureate Yann LeCun stated he has "zero concerns" about AI wiping out humanity and attributes recent "rogue" AI incidents to poor human oversight and system design. LeCun also criticized the effective altruism movement, calling it "super toxic" and a "complete disaster" for its focus on AI's existential risks.
OpenAI's GPT-6 Astra AI Clears WoW Orc Starting Zone Using Network Data
OpenAI's GPT-6 Astra AI model successfully completed the Orc starting zone in World of Warcraft in 40 minutes with no deaths. The AI operated by interpreting network traffic and quest data directly from server files, rather than visual input, demonstrating a novel approach to AI interaction with complex game environments.
Sean Parker Reorients Stability AI Towards Music Generation with Major Label Backing
Sean Parker and CEO Prem Akkaraju are repositioning Stability AI to focus on AI tools for music professionals. The company secured $76 million in funding and licensing deals from Sony, Warner, and Universal, and has released new audio models and music-editing software.
Wall Street Banks Increase AI Job Postings by 49%, Demand for AI Agent Skills Jumps 1,721%
Job postings for AI-related roles at major banks like JPMorgan Chase, Citigroup, and Capital One increased by 49% this year, totaling 139,819 listings. The demand for skills in AI agent orchestration saw a 1,721% rise, indicating banks are moving beyond chatbots to integrate AI agents into core business operations.
Circuit Breaker Labs develops AI agents to test models for psychologically harmful interactions
Circuit Breaker Labs, a Startup Battlefield 200 finalist, created AI agents that simulate diverse users to test AI models for detecting dangerous and psychologically harmful interactions. This initiative addresses concerns over AI models contributing to user harm, including wrongful death lawsuits against Character.AI and OpenAI.
Study identifies specific linguistic tells in AI-generated text from Claude and GPT models
A study by Graphite analyzed AI writing habits, identifying over 13,000 phrases used significantly more often by AI models than humans. The research found that Claude models are converging with human word distribution, while GPT models are diverging, with each model exhibiting unique linguistic patterns.
Shopify Launches Canvas AI Site Builder for E-commerce Stores
Shopify introduced Canvas, a new AI-powered tool that allows merchants to build and customize their online stores by chatting with the AI agent Sidekick. This development simplifies store creation by enabling real-time visual updates through conversational commands, reducing the need for manual coding or developer intervention. It matters because it represents a shift towards AI-driven no-code website development, making e-commerce store setup more accessible and efficient for merchants.
Ataraxos AI defeats world's best Stratego player, overcoming hidden information challenge
Researchers from Carnegie Mellon, MIT, NYU, and Stanford developed Ataraxos, an AI that defeated the top Stratego player, Pim Niemeijer, 15-1 with four draws. This achievement is significant because Stratego is an imperfect-information game with a large number of hidden variables and long game lengths, which previously challenged AI systems like DeepMind's DeepNash.
Huawei updates AI accelerator roadmap, accelerates Ascend 960, introduces 970/980
Huawei updated its AI hardware roadmap at Huawei Connect, accelerating the Ascend 960 timeline and introducing Ascend 970 and 980 specifications. The company is transitioning to new SIMD+SIMT architectures for improved AI workload performance, despite current production capacity limitations.
Consultancy tests DeepSeek V4.1 Flash on Nvidia H200s, finds it more expensive than Claude
The Call Center Doctors rented Nvidia H200 hardware to test DeepSeek V4.1 Flash, finding that running it on rented hardware was more expensive and less efficient than using Claude's API for their specific coding agent workload. The test aimed to verify claims of DeepSeek being significantly cheaper than Claude, but revealed that high re-read token volume made the hardware-based solution impractical.
AI Chatbots Remove Hijabs from Images of Muslim Women When Prompted
AI chatbots like ChatGPT and Grok removed hijabs from images of Muslim women when prompted, while Gemini did so when asked to make a woman look more "western." This behavior raises concerns about AI's role in perpetuating anti-Muslim sentiment and digital alteration without consent.
Bank of England Governor Warns AI Boom Could Trigger Financial Market Shocks
Andrew Bailey, Governor of the Bank of England, warned that the significant investment in AI could lead to financial market shocks and a correction in asset prices. He stated that while AI offers growth potential, the current market valuation assumes all companies will succeed, which is unlikely.
Anthropic's Claude AI Used in Malaysian Election Influence Operation
Anthropic revealed its Claude AI was used in a political influence operation during recent Malaysian state elections, creating fake news outlets and social media accounts to spread disinformation. This incident highlights how AI can simplify and scale disinformation campaigns, posing a significant risk to future elections, particularly in regions vulnerable to fake news like Southeast Asia.
OpenAI CEO Sam Altman declined to testify at Senate hearing on AI risks
OpenAI CEO Sam Altman declined an invitation to testify at a Senate subcommittee hearing on the risks of "rogue AI." Senator Josh Hawley, who chairs the subpanel, stated Altman's refusal was "unfortunate" given recent incidents involving OpenAI models and the need for transparency regarding AI development.
Brain's Neural Waves Exhibit Complex Patterns Beyond Simple Oscillations
New research indicates that neural waves in the brain are more complex than previously understood, exhibiting source, sink, and spiral patterns. These complex wave patterns are associated with different behavioral tasks and may facilitate real-time brain reorganization.
Cohere Releases Embed 5 with Pro and Fast Models for Optimized RAG Workloads
Cohere launched Embed 5, introducing Pro and Fast models that share an embedding space, allowing users to index data with the Pro model and query with the faster, cheaper Fast model. This setup aims to reduce latency and cost in RAG and agent workloads, with Cohere's tests showing minimal impact on retrieval quality.
Google DeepMind Develops Watermarking for AI-Designed Proteins
Google's DeepMind team published research on a protein watermarking system to identify AI-designed proteins. This system embeds a watermark directly into protein sequences without affecting their function, allowing for the distinction between trusted and potentially threatening AI-generated proteins.
ElevenLabs Reportedly Valued at $22 Billion, Reaches $600M ARR
ElevenLabs, a company specializing in AI voice models, has reportedly achieved a $22 billion valuation from investors and reached $600 million in annual recurring revenue (ARR). The company's technology is used by major enterprises for customer service and by creators for audiobooks and dubbing.
Lawyer Uses ChatGPT to Generate Fake Witnesses and Testimony in Murder Appeal
A New Mexico attorney submitted a legal brief containing fabricated witnesses and testimony generated by ChatGPT in a murder appeal case. The New Mexico Supreme Court issued a direct contempt order, highlighting the risks of relying on AI for legal work without verification.
OpenAI's Dots Autonomous Agents Utilize Nine-Core AMD EPYC VMs
Geekbench 7 results indicate that OpenAI's newly launched Dots autonomous agents run on virtual machines powered by nine-core AMD EPYC 9V74 processors with 9.73GB of memory. This hardware configuration provides significantly higher multi-core performance compared to Meta's Muse agents, which use two-core AMD EPYC Turin processors.
SynthID Bio Introduces Watermarking for AI-Generated Biological Designs
SynthID Bio has launched a watermarking approach for AI-generated biological designs to strengthen biosecurity and information integrity. This technology embeds a verification layer into biological designs, helping to track provenance and streamline DNA synthesis screening.
Microsoft Fabric Integrates AI Agents for Enterprise Data Understanding
Microsoft is positioning its Fabric data platform as a central hub for AI agents to understand enterprise operations. This integration allows agents to access business context, with Fabric IQ now feeding Microsoft 365 Copilot and Power BI gaining capabilities to convert definitions into applications.
Pearson to Acquire AI Skills Platform Workera in H2 2026
Pearson will acquire Workera, an AI-native skills platform, in the second half of 2026. This acquisition aims to strengthen Pearson's enterprise offerings in adaptive skills training and assessments to address the growing demand for AI workforce upskilling.
Open TTS Leaderboard Launches for Scalable, Objective Text-to-Speech Evaluation
A new Open TTS Leaderboard has been introduced to provide scalable, objective evaluation for text-to-speech and voice cloning models. It uses metrics like word error rate, inference speed, and speaker similarity to address the limitations of human preference-based arena leaderboards, which struggle with scalability and open-source model representation.
AI-generated images raise concerns about election integrity and misinformation
The increasing ease of creating AI-generated images is causing alarm regarding their potential to spread misinformation and influence elections. Recent incidents involving fake images of public figures highlight the challenge of distinguishing truth from falsehood, impacting political discourse and public trust.