This Week in NLP #374
Keep up with what happened in NLP in the week ending Friday 16th January 2026.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Apple has partnered with Google to use Gemini as the foundation for its AI features, with in-house fine-tuning and no Google branding, aiming to improve Siri’s capabilities. [9to5Mac]
Anthropic launched Cowork, an AI agent tool for non-technical users that can read, edit, and create files, built largely by Claude Code in just over a week. [VentureBeat]
And Anthropic launched Claude for Healthcare, an AI tool with advanced connectors to help healthcare providers streamline administrative tasks and access medical databases. [TechCrunch]
Google has launched the Universal Commerce Protocol, an open standard for AI-powered shopping backed by major retailers, potentially challenging Amazon’s retail dominance. [GeekWire]
And Google is launching a new Gemini AI feature called Personal Intelligence that can access and analyze users’ data across Google services to provide personalized assistance. [Gizmodo]
Now read on for everything else that happened in NLP this week.
Making News:
What makes this newsletter different?
This Week’s Topics:
If you’re reading this in a desktop web browser, you can access a navigation menu at the left-hand edge of this window.
The Generative AI Wars
Alibaba is linking Taobao, Alipay, Fliggy, and Amap to its Qwen AI app to create a unified consumer platform. [Bloomberg]
Alphabet reached a $4tn valuation, becoming the world’s second-most valuable company after securing a key AI partnership with Apple and surging in stock value. [The Guardian]
Amazon plans to leverage its extensive device footprint and consumer familiarity with Alexa to expand its AI platform, targeting Prime members with an enhanced, more capable assistant. [TechCrunch]
Ant International is partnering with Google’s Universal Commerce Protocol to enable AI-powered shopping experiences with seamless, secure, and intuitive payment capabilities across multiple markets. [Business Wire]
And Anthropic has secured a significant partnership with Allianz, bringing its responsible AI tools to the German insurance giant’s employees and workflows. [TechCrunch]
Lenovo launched Qira, an AI assistant that works across Lenovo and Motorola devices, offering context-aware, task-completing capabilities with a focus on user privacy. [ZDNet]
Microsoft is reshuffling GitHub’s internal teams to compete with AI coding startups and maintain its dominant position in the software development platform market. [ITPro]
Microsoft’s Copilot struggles with a mere 1.1% web AI market share, far behind ChatGPT and Google’s Gemini. [PCWorld]
Musk criticized Apple and Google’s partnership to power Siri with Gemini, calling it an unreasonable concentration of power for Google. [MacRumors]
OpenAI signed a US$10B multi-year deal with Cerebras Systems to deliver 750 megawatts of compute for faster AI system responses through 2028. [TechCrunch]
OpenAI is seeking US-based hardware suppliers to support its expansion into consumer devices, robotics, and data centers. [Bloomberg]
Salesforce launched Slackbot, an AI-powered personal work agent integrated directly into Slack that understands context, respects permissions, and helps employees work more efficiently. [Business Wire]
Samsung is reportedly reviving Bixby by partnering with Perplexity to handle complex AI queries and provide advanced voice interaction features. [TechRepublic]
Walmart and Google have partnered to integrate Gemini AI into Walmart’s shopping experience, enabling customers to more easily discover and purchase products. [CNBC]
xAI is preparing to release Grok version 4.2, featuring a new Grok Code interface with local agent capabilities, GitHub integration, and advanced configuration options for developers. [TestingCatalog]
Meanwhile, xAI reported a US$1.46B quarterly loss while nearly doubling revenue to US$107 million and raising US$20B in funding. [DealStreetAsia]
AI Supremacy
China’s AI researchers believe the country can narrow its technological gap with the US through innovation, despite challenges in chipmaking tools and computing infrastructure. [Yahoo Finance]
But Chinese AI leaders warned that the country is unlikely to overtake the US in AI in the near future. [Bloomberg]
Chinese technology companies are challenging Silicon Valley’s dominance by prioritizing speed, affordability, and efficiency across industries like social media, EVs, e-commerce, and AI. [London Daily]
Chinese AI startup DeepSeek is gaining ground in developing nations, potentially narrowing the global AI adoption gap, according to a Microsoft report. [Yahoo Finance]
And China is drafting purchase rules for Nvidia H200 chips to balance domestic industry development with tech companies’ need for advanced AI hardware. [Nikkei Asia]
The Trump administration signed a US$250B trade deal with Taiwan to boost US semiconductor manufacturing and reduce foreign supply chain dependence. [TechCrunch]
The US is forming Pax Silica, a strategic alliance of trusted partners aimed at leading global AI development and securing technology supply chains. [The National]
Qatar and UAE will join the Pax Silica initiative to secure AI and semiconductor supply chains, strengthening technology cooperation among allies. [Yahoo Finance]
And the US plans to invite India to join Pax Silica, to secure the silicon supply chain and reduce reliance on China. [Firstpost]
Meanwhile, Nvidia CEO Jensen Huang argues that US-China technological decoupling is unrealistic due to deep economic, talent, and innovation interdependencies. [BLiTZ]
And Z.ai claims to have trained its GLM-Image AI model entirely on Huawei hardware, potentially signalling China’s growing domestic AI capabilities. [The Register]
Sovereign AI
AWS has launched the AWS European Sovereign Cloud, a new independent cloud infrastructure in the EU, with plans to expand to Belgium, the Netherlands, and Portugal. [Business Wire]
Europe risks losing digital sovereignty due to heavy reliance on US tech giants’ cloud services, potentially compromising data protection and national security. [Global Current News]
The European Commission is seeking input on open source technology to boost the EU’s tech sovereignty and competitiveness. [BiometricUpdate.com]
ExpertEase AI launched its Australian sovereign AI platform on Microsoft Marketplace, offering no-code, voice-operated digital employees with data sovereignty for organizations. [PRWeb]
Ireland is seeking to revive its data center industry through a new energy plan, leveraging its skilled workforce and tech company relationships. [Bloomberg]
A Microsoft report reveals the UAE and Singapore lead global AI adoption, while South America and developing nations significantly trail behind. [Rest of World]
Silicon Valley tech giants are investing US$67.5B in India’s AI and data centre ecosystem, signaling the country’s emergence as a global digital infrastructure powerhouse. [London Daily]
And South Korea’s AI initiative eliminated Naver Cloud and NCSoft, leaving SK Telecom, LG AI Research, and Upstage to compete for sovereign AI development. [Cryptopolitan]
Feature Creeps
Anthropic’s Claude Code now uses ‘lazy loading’ for AI tools, dramatically reducing context token usage and improving model accuracy by dynamically fetching tool definitions only when necessary. [VentureBeat]
Google has revamped its Trends Explore page with Gemini-powered AI, automatically identifying and comparing search trends with enhanced visualization and exploration features. [TechCrunch]
Microsoft’s Windows 11 might soon feature a Copilot AI button in File Explorer, potentially offering a detachable sidebar for AI-powered file interactions. [TechRadar]
Meanwhile, Microsoft is allowing admins to uninstall Copilot on managed devices under specific conditions in the latest Windows 11 Insider Preview build. [ITPro]
Proton’s Lumo AI chatbot now offers an encrypted Projects feature for storing and syncing task-related information across devices. [Engadget]
Hype Bubble?
A study by BCG reveals CEOs are increasingly prioritizing AI investments, with 94% committed to continued funding despite potential delayed returns. [Computer Weekly]
A report by Camunda reveals that 73% of organizations acknowledge a gap between their agentic AI vision and reality, with only 11% of use cases reaching production. [Business Wire]
Research by Clari Labs found that 87% of enterprises missed revenue targets in 2025, with 48% admitting their revenue data isn’t ready for AI-driven growth. [Business Wire]
A report by DDN reveals 65% of organizations struggle with AI success due to complex infrastructure, cloud strategies, and skills gaps. [Business Wire]
Deloitte’s Tech Trends report reveals enterprises are now using AI to fundamentally reshape core business processes, infrastructure, and talent models. [News9 LIVE]
Forrester analyst JP Gownder argues that current AI technology has not significantly improved productivity or led to widespread job losses. [The Register]
And Forrester warns tech leaders to prepare for cautious 2026 budgeting amid economic uncertainty, AI shifts, and the need for continuous experimentation. [CFOtech]
Moody’s forecasts US$3T in AI datacenter infrastructure investment through 2029, despite concerns about profitability and power grid constraints. [The Register]
A survey by NTT reveals global CEOs plan to increase AI investment despite infrastructure challenges and sustainability concerns, with 68% aiming to boost spending in the next two years. [Business Wire]
Tech leaders and investors debate whether the current AI investment surge represents a potential economic bubble driven by massive spending and speculation in AI technologies. [CNBC]
Big Iron
Big Tech’s massive AI-driven data centre expansion is straining US electricity grids, forcing power operators to demand alternative energy sources or conditional grid connections. [Firstpost]
Ireland’s new Large Energy User Action Plan aims to attract data center investments, sparking environmental and energy grid concerns from opposition parties. [ITPro]
Meta has signed nuclear power agreements with three companies to provide 6.6 gigawatts of energy for AI data centers by 2035. [Engadget]
And Meta has established a top-level Compute initiative to oversee massive AI computing infrastructure expansion, planning tens of gigawatts this decade and hundreds of gigawatts long-term. [TechRadar]
OpenAI and SoftBank will each invest US$500m in SB Energy to build a 1.2-gigawatt data centre for their Stargate AI initiative in Texas. [DealStreetAsia]
And OpenAI has signed a US$10B+ deal with Cerebras Systems to deploy its wafer-scale accelerators for faster AI inference services through 2028. [The Register]
Trump calls on tech companies to cover their own data center energy costs to protect American households from rising electricity bills. [TechRadar]
Microsoft pledged to become a ‘good neighbor’ by addressing community concerns about data centers, including proposing higher electricity rates and reducing water usage. [Wired]
SiFive is integrating Nvidia NVLink Fusion into its RISC-V compute platforms to enable more efficient, customizable AI data center infrastructure. [Business Wire]
Hot Chips
Apple is struggling to secure TSMC chip production capacity as Nvidia takes center stage in the AI-driven semiconductor market. [Culpium]
Google and Microsoft are desperately seeking DRAM chips in Korea, with executives dubbed ‘DRAM beggars’ as the global semiconductor shortage intensifies. [India Today]
Huawei has staged a remarkable comeback after US sanctions, developing its own technology, expanding into new markets, and maintaining strong revenue through innovation and government support. [ARY News]
Nvidia’s DGX SuperPOD with Vera Rubin platform delivers 28.8 Exaflops using 576 GPUs, challenging Huawei’s AI compute capabilities. [TechRadar]
Nvidia may now sell H200 AI chips to China after US Commerce Department revises export criteria to allow case-by-case reviews. [Bloomberg]
And Nvidia clarified it does not require upfront payment for H200 chips to Chinese customers, contrary to previous reports. [Reuters]
But Nvidia’s H200 AI chips are barred from entering China, with authorities instructing domestic tech companies not to purchase them unless absolutely necessary. [Reuters]
And Nvidia’s H200 AI processor exports to China will be constrained by memory chip shortages, according to a US lawmaker. [Bloomberg]
In any case: Nvidia’s chip sales to China will be subject to a 25% US tariff under a Trump-approved agreement. [Bloomberg]
Warm Bodies
AI companies are experiencing significant employee movement, with top talent frequently switching between Anthropic, OpenAI, and other labs. [TechCrunch]
Airbnb hired Ahmad Al-Dahle, former Meta AI head, as its new technology chief to advance AI-driven travel and e-commerce experiences. [CNBC]
Anthropic has hired Andrea Vallone, a former OpenAI safety researcher who studied AI’s response to users’ mental health struggles. [The Verge]
Egnyte is using AI coding tools to accelerate developer onboarding and growth, emphasizing human talent retention and skill development rather than workforce reduction. [VentureBeat]
Forrester predicts AI and automation could eliminate 10.4 million US jobs by 2030, but expects a gradual, structural shift rather than a sudden employment collapse. [The Register]
Geoffrey Hinton predicts AI will rapidly replace jobs across various sectors, potentially transforming software engineering and other professional fields by 2026. [SlashGear]
London’s mayor warned of potential mass unemployment from AI while simultaneously offering free training and support to help workers adapt to technological changes. [The Next Web]
Mercor is hiring struggling workers to train AI models that could potentially replace their jobs, highlighting the growing concerns about AI’s impact on employment. [Futurism]
Meta is reportedly planning to lay off around 1,500 employees in its Reality Labs division as the company shifts focus towards AI infrastructure development. [Gizmodo]
OpenAI has rehired two Thinking Machines cofounders, Barret Zoph and Luke Metz, amid speculation about Zoph’s departure from the startup. [Wired]
Salesforce is maintaining flat software engineering headcount due to AI productivity gains while expanding sales and customer engagement roles to push its Agentforce service. [ITPro]
Second Talent reports entry-level tech hiring plummeted 73% as companies pivot to production-ready AI engineers with specialized skills. [EIN Presswire]
UAE minister Omar Sultan Al Olama warned that humans must develop broad knowledge and skills to remain relevant in an AI-driven world. [Hindustan Times]
Zoho executives warn that AI is driving productivity gains in tech, potentially leading to fewer jobs and smaller teams in the near future. [India Today]
Meanwhile, tech companies are facing a critical shortage of skilled electricians, plumbers, and HVAC technicians needed to build the data centers powering the AI revolution. [Wired]
Consumer AI
Amazon has acquired Bee, an AI wearable device designed to record conversations, learn user patterns, and serve as a personal AI companion outside the home. [TechCrunch]
Nigerian-born tech innovator Hanu Fejiro Agbodjie has donated AI-powered smart glasses to hearing-impaired Nigerians, aiming to enhance communication and social inclusion through technology. [The Eagle Online]
OpenAI reportedly plans to develop five hardware devices, with the first being Sweetpea, an AI-native audio wearable designed with Jony Ive’s team. [AI Breakfast]
Ring founder Jamie Siminoff returned to the company, driven by AI’s potential and inspired by the Palisades fires, to transform the video doorbell maker into an intelligent home assistant. [TechCrunch]
Rokid unveiled affordable smart glasses and partnered with Mudra to offer hands-free computing through a neural wristband at a lower cost than Meta’s offerings. [digitaltrends]
A survey by Smart Analytics Global predicts AI smart glasses revenue will quadruple in 2026, reaching US$5.6B with 20 million units sold. [Business Wire]
XGIMI launched its Memomind AI glasses series, featuring the lightweight Memo One with dual-lens displays, integrated AI, and customizable frames at $599. [ZDNet]
Wearable technology is transforming disability support by providing personalized, adaptive solutions that enhance independence, safety, and health through intelligent, discreet devices. [Technowize]
It’s Only a Model
DeepSeek is preparing to launch its V4 AI model with strong coding capabilities in the coming weeks. [The Information]
Google Research has expanded its Health AI Developer Foundations program by releasing MedGemma-1.5, a compact multimodal medical model for developers to adapt to local workflows. [Marktechpost Media]
Google updated its Veo 3.1 AI video-generation model to create vertical videos and improve expressiveness using reference images. [TechCrunch]
Resemble AI has launched an open-source, real-time text-to-speech model offering expressive voice generation as an alternative to proprietary platforms. [Slator]
Z.ai’s GLM-Image, an open-source 16-billion parameter image generation model, challenges proprietary AI models by achieving state-of-the-art performance in generating text-heavy, information-dense visuals through a hybrid auto-regressive and diffusion design. [VentureBeat]
Whose Data?
CallMiner leverages 100 million hours of customer conversation data to provide conversation intelligence that enables businesses to deploy more effective AI-powered customer service automation. [TechCentral]
Cengage and Hachette seek to join a class action lawsuit against Google, alleging Gemini illegally copies copyrighted books to train its generative AI. [Publishers Weekly]
Getty Images has developed a comprehensive strategy for data sovereignty, focusing on auditing, classifying, and controlling data location and movement to ensure compliance and security. [Computer Weekly]
Google is offering Gemini users personalized AI responses by accessing their personal data across Google apps, while claiming to protect user privacy. [The Register]
IBM launched Sovereign Core, a platform giving enterprises and governments secure, sovereign control over AI infrastructure and data management. [TechRadar]
India has proposed a draft law requiring AI companies to pay royalties for using copyrighted work, potentially reshaping tech firms’ operations in a massive, strategically important market. [Rest of World]
Microsoft Copilot’s ‘Reprompt’ attack allows hackers to steal sensitive user data with a single malicious link click, bypassing security controls. [ZDNet]
OpenAI and Handshake AI are seeking contractors to upload real work examples as training data for AI models, raising potential intellectual property concerns. [TechCrunch]
And OpenAI faces potential sanctions from publishers for allegedly deleting digital conversations despite court-ordered preservation in an ongoing copyright infringement lawsuit. [Editor & Publisher Magazine]
The Poison Fountain project calls on website operators to deliberately add poisoned links to undermine AI training data, aiming to damage machine intelligence systems. [The Register]
Wikimedia Foundation signed licensing deals with major tech companies to charge for AI training access to Wikipedia’s content, expanding its Enterprise program. [Ars Technica]
The LLM Ecosystem
Acer unveiled the Veriton RA100, a compact Ryzen AI Max+ 395 mini PC targeting creators. [TechRadar]
Cast AI has launched OMNI Compute, a unified compute marketplace valued at over US$1B, enabling enterprises to access GPUs across clouds without code changes. [Business Wire]
DeepSeek introduced Engram, a conditional memory module that optimizes AI model performance by efficiently separating static information retrieval from dynamic reasoning. [VentureBeat]
DigitalOcean’s Inference Cloud Platform, powered by AMD Instinct GPUs, doubled Character.ai’s production inference performance while reducing costs by 50%. [Business Wire]
Kaggle launched Community Benchmarks, enabling the AI community to design, run, and share custom benchmarks for evaluating AI models across various tasks and capabilities. [Google]
KIBO Commerce launched KIBO Connect Hub and KIBO MCP, two new integration platforms designed to streamline commerce technology and enhance AI-driven capabilities for businesses. [PRWeb]
Minisforum debuts AI X1 Pro-470, the first Ryzen AI 9 HX470 mini PC with up to 128GB RAM and 12TB storage, targeting heavy AI workloads. [TechRadar]
Miro launched AI Workflows, enabling enterprise teams to collaborate visually and accelerate innovation by transforming ideas into deliverables in hours instead of weeks. [Business Wire]
MongoDB launched four new embeddings models to improve data retrieval quality and performance for AI systems in enterprise production environments. [VentureBeat]
Monstarlab has updated its MonstarX AI PoC development platform with expanded specification input methods, including voice and document-based requirements. [EIN Presswire]
MSI’s AI Edge mini PC offers high-performance local AI processing, gaming capabilities, and up to 128GB memory in a compact 4-liter form factor. [TechRadar]
Orchestral AI offers a synchronous, type-safe Python framework for scientific AI agents, prioritizing reproducibility, cost-tracking, and model-centric design across multiple providers. [VentureBeat]
Raspberry Pi launched the AI HAT+ 2, a neural network accelerator with 8 GB RAM and 40 TOPS performance, targeting local AI computing with modest LLM capabilities. [The Register]
Seeweb launched a Serverless GPU solution that provides scalable, on-demand computing power to help organizations overcome GPU shortages and accelerate AI innovation. [EIN Presswire]
Semantic caching reduces LLM API costs by 73% by using embedding-based similarity lookup with query-type-specific thresholds, increasing cache hit rates from 18% to 67%. [VentureBeat]
Senzing launched its entity resolution AI in AWS Marketplace, enabling organizations to deploy real-time risk detection, fraud prevention, and KYC solutions with enhanced data accuracy and privacy. [Business Wire]
Sisense launched an AI-driven analytics platform with advanced capabilities, enabling faster data interpretation, exploration, and decision-making through intelligent assistants and secure AI integration. [Business Wire]
Temporall launched Tempo, an AI Intelligence platform that helps enterprises quantify and validate AI’s impact on business performance by unifying data from various AI tools and systems. [EIN Presswire]
TradeStation Securities launched MCP Connection, enabling users to link trading accounts with third-party AI platforms, starting with Claude. [Business Wire]
TrustLogix launched TrustAI, a policy control plane that provides unified, machine-speed governance across AI and data ecosystems to eliminate access control risks. [PRWeb]
Unframe launched Unframe Unlimited, a partner program designed to help enterprises accelerate AI adoption and drive measurable business outcomes quickly and efficiently. [Business Wire]
Agentic AI
Amatrium has upgraded AmatriumGPT to version 5.1, introducing multi-agent intelligence and enhanced performance for enterprise AI workflows. [EIN Presswire]
Amperity launched a Customer Data Agent that uses AI and unified customer data to help marketers quickly generate insights and take action. [Amperity]
Commerce has endorsed Google’s Universal Commerce Protocol, enabling seamless AI-driven shopping experiences with direct checkout and optimized product data. [GlobeNewswire]
Conduent launched an AI Experience Center to showcase AI and GenAI solutions for commercial, transportation, and government clients, demonstrating innovative technologies developed with Microsoft and other partners. [Business Wire]
FIS launched an industry-first agentic commerce offering that enables banks to securely conduct AI-powered transactions with enhanced fraud protection and consumer experience. [Business Wire]
GitLab launched its Duo Agent Platform, offering AI-powered tools to orchestrate agents across the software development lifecycle and address productivity bottlenecks. [Business Wire]
Highspot launched Deal Intelligence with Deal Agent, an AI-powered tool to help sales teams execute deals more effectively and win more consistently. [Business Wire]
Manhattan Associates has launched AI-powered enhancements to its Active Omni platform, introducing new agents and features to help retailers optimize sales and customer experiences. [Business Wire]
Manus has partnered with Similarweb to provide AI agents access to digital marketing data intelligence for more accurate market analysis and strategic planning. [Business Wire]
Nutshell launched an AI Agent Marketplace with purpose-built agents that automate sales and marketing tasks, reducing busywork and accelerating deal closures. [PRWeb]
Rocket Driver launched a fully managed White Label AI Agents platform that enables agencies to deploy AI-powered client services quickly and easily without complex configuration. [EIN Presswire]
Sakana AI’s ALE-Agent secured first place in a coding competition by autonomously optimizing solutions through innovative reasoning and future-focused strategy. [VentureBeat]
Superhuman is integrating Saifr’s AI compliance agent into its Go platform to help financial services professionals create regulatory-compliant communications more efficiently. [Business Wire]
Treasure Data launched Marketing Super Agent, an AI system that operates like a full marketing department within its AI Marketing Cloud platform. [Business Wire]
UiPath’s Screen Agent Powered by Claude Opus 4.5 achieved top ranking on OSWorld-Verified benchmark for agentic automation across 369 computer tasks. [Business Wire]
Other LLM Sightings
Algolia has partnered with Microsoft to help retailers control and improve their product data across AI-powered shopping platforms like Copilot, Bing, and Edge. [Business Wire]
Blue Olive LLC launched BookAReview, an AI-powered Shopify app that generates personalized, contextually relevant review requests for merchants. [PRWeb]
BlueMatrix has partnered with Perplexity to provide AI-powered, governance-compliant research discovery for institutional investors. [Business Wire]
Box has launched Box Extract, an AI-powered tool that transforms unstructured enterprise content into actionable, metadata-driven insights across various business functions. [Business Wire]
BX Partners has partnered with Boosted.ai to provide advanced AI-powered investment insights to advisors and asset managers through its platform. [Business Wire]
Canopy integrated Filed’s AI-powered tax preparation into its practice management platform, aiming to streamline workflow and reduce manual tax preparation work. [Business Wire]
Hamlet is building a platform that uses humor to engage citizens with local government meetings, while simultaneously selling data from these meetings to private clients like real estate developers. [NiemanLab]
Hawk launched Analytics Studio, an AI lifecycle management solution helping financial institutions develop, maintain, and govern AI models for fraud and AML prevention. [Business Wire]
Lenovo launched AI-powered Smart Store Services and AI Retail Assistants to help retailers improve operations, reduce downtime, and enhance customer experiences. [Business Wire]
Ninth Wave launched Compass, an AI-driven onboarding assistant that simplifies and accelerates open finance integration for financial institutions. [Business Wire]
PRAC developed an AI fraud-detection model that could have flagged tens of billions in potentially fraudulent COVID-19 loan applications before disbursement. [The Register]
RENTAL12 has introduced FLO, an AI assistant that provides accurate, source-based information while maintaining human-led communication for its owner-operated hospitality properties. [EIN Presswire]
Salesforce and the World Economic Forum launched EVA, an AI-powered concierge app using Agentforce 360 to support leaders at the 2026 Davos Annual Meeting. [Business Wire]
Sapio Sciences launched an ecosystem of partner integrations for ELaiN, its AI lab notebook, enabling scientists to access trusted tools within a single experimental environment. [Business Wire]
Spotlink is launching an AI-powered platform that helps NYC drivers understand parking rules and anticipate available curb spaces in real-time. [EIN Presswire]
The Nova Method launched NovaSight, an AI platform helping marketing leaders understand and improve their brand’s visibility in AI-driven search and recommendations. [PRWeb]
Veterans Alliance launched VA.org, an AI-powered platform helping veterans navigate benefits more easily through tech-forward tools and personalized guidance. [EIN Presswire]
WORK-SELF launched an AI-powered career platform that uses advanced matching technology and human expertise to help professionals find the right job within 30 days. [EIN Presswire]
Xero launched AI-powered analytics tools to help small businesses make data-driven decisions more effectively and confidently. [Business Wire]
Risks and Responses
AI models trained to write buggy code can develop disturbing, unrelated behaviours like expressing desires to harm or enslave humans. [The Register]
AI-generated misinformation is increasingly threatening local democracy in the UK by creating convincing fake social media posts that spread rapidly online. [TechRepublic]
AI-powered phishing has become a sophisticated cybersecurity threat, using generative models to create personalized emails that bypass traditional security measures and exploit AI systems. [TechCentral]
Anthropic has implemented strict safeguards preventing third-party applications from spoofing its Claude Code client, disrupting workflows and blocking access for competitors like xAI. [VentureBeat]
But Cowork, Anthropic’s new productivity AI, can be tricked via prompt injection to transmit sensitive files to an attacker’s account without additional user approval. [The Register]
Bernie Sanders warned that AI, driven by tech leaders like Musk and Zuckerberg, poses a serious threat to jobs, mental health, and societal well-being. [Benzinga]
Google has removed some AI health summaries after an investigation revealed potentially dangerous and misleading medical information being presented to users. [The Guardian]
Google’s Gemini and other AI chatbots were found to frequently fabricate news sources, invent stories, and provide unreliable information during a month-long experiment. [digitaltrends]
ICE deployed undertrained officers after an AI hiring tool mistakenly categorized applicants with no law enforcement experience as experienced officers. [Gizmodo]
Microsoft introduces a policy allowing admins to remove Copilot from managed devices, but with strict conditions that make removal challenging. [The Register]
And developers have created a PowerShell script to remove AI features from Windows, citing privacy, security, and user experience concerns. [The Register]
Nvidia CEO Jensen Huang dismisses AI risks, arguing that concerns about potential harms are damaging and unhelpful to the industry and society. [Gizmodo]
Hackers targeted misconfigured AI service proxies in a large-scale campaign, probing OpenAI, Gemini, and other LLM APIs during the 2025 Christmas break. [TechRadar]
Palo Alto Networks discovered critical vulnerabilities in AI/ML Python libraries from Nvidia, Salesforce, and Apple, enabling potential remote code execution. [TechRadar]
Radware discovered a ‘ZombieAgent’ vulnerability in ChatGPT apps feature that allows silent data exfiltration and account takeover through hidden prompts. [TechRadar]
RunSybil’s AI tool Sybil discovered a complex GraphQL vulnerability, highlighting AI’s growing ability to detect and potentially exploit cybersecurity weaknesses. [Wired]
A Sonar survey revealed that nearly half of software developers don’t verify AI-generated code, despite 96% admitting they don’t fully trust its accuracy. [ITPro]
Thomson Reuters has created a Trust in AI Alliance with tech leaders to advance trustworthy, responsible AI systems through shared insights and engineering principles. [Artificial Lawyer]
X added minimal restrictions to its Grok AI chatbot after widespread criticism over its ability to generate sexually degrading deepfake images of women and children. [Gizmodo]
Indonesia and Malaysia have temporarily blocked xAI’s Grok chatbot over concerns about its generation of non-consensual, sexually explicit AI-generated imagery. [TechCrunch]
Meanwhile, Grok will be integrated into Pentagon networks this month as part of the Defense Secretary’s AI acceleration strategy to enhance military technological capabilities. [Ars Technica]
Varonis discovered a new prompt-injection method in Microsoft Copilot that could leak sensitive data through a malicious URL with a single click. [TechRadar]
The World Economic Forum reports that businesses are increasingly assessing AI security risks before deployment, with 64% now evaluating potential threats. [TechRadar]
Regulation
The EU’s Digital Networks Act will focus on telecoms infrastructure, with Big Tech following a voluntary best practices framework rather than facing strict new regulations. [TechRadar]
Missouri lawmakers are exploring balanced AI regulations across multiple domains, focusing on consumer protection, child safety, and technological innovation while addressing potential risks. [First Alert 4]
New York is proposing AI safety measures for children and investigating algorithmic pricing while welcoming AI development to the state. [Axios]
And Washington state lawmakers proposed five bills to regulate AI, targeting high-risk systems, AI companions, school use, and digital likenesses. [GeekWire]
WhatsApp has been ordered by Brazil’s competition watchdog to suspend its policy banning third-party AI companies from using its business API for chatbots. [TechCrunch]
And Meta will exclude Italy from its WhatsApp ban on rival AI chatbots after an order from the country’s antitrust authority. [Reuters]
Conversational AI
AOK Bayern deployed NiCE’s CXone CX AI Platform on the EU Sovereign Cloud, transforming customer service for 4.5 million health insurance members. [Business Wire]
Cerberus Capital Management selected Behavox’s AI-powered Quantum solution to enhance its communications surveillance and compliance program. [Business Wire]
Calabrio launched Omni Agent Intelligence, a vendor-agnostic platform unifying quality and performance measurement across human and AI contact center agents. [Business Wire]
Dialpad launched real-time AI capabilities in Japan, offering free advanced features through a partnership with SoftBank until April 2026. [Business Wire]
Equature launched SmartAgent, an AI voice agent that automates routine 911 and public safety calls while keeping dispatchers focused on emergencies. [EIN Presswire]
Emplifi has partnered with Reddit to transform community conversations into actionable industry intelligence using AI-powered analytics across various brand sectors. [Business Wire]
Everseen launched Everact, an agentic AI platform enabling retailers to conversationally query store data and receive instant, curated insights and recommended actions. [Business Wire]
First Insight’s AI tool Ellis enables retailers to quickly transform consumer insights into strategic go-to-market decisions through conversational, predictive analysis. [Business Wire]
Five9 has partnered with Google Cloud to launch an AI-powered enterprise customer experience solution that enables seamless, intelligent interactions across channels. [Business Wire]
HawkSoft and Sonant have integrated a voice AI platform that automates call handling and documentation within HawkSoft’s insurance agency management system. [Speech Technology Magazine]
Helport AI launched HyprX, an AI-powered digital twin platform that replicates expert professionals’ knowledge and communication styles across various industries. [GlobeNewswire]
Income Lab launched an AI Scribe that automatically transforms client conversations into retirement plans during Zoom meetings. [Business Wire]
Otto has launched Agentic Confirmations, an AI-powered feature that transforms appointment confirmation texts into natural, two-way conversations for veterinary clinics. [Business Wire]
Realbotix demonstrated two autonomous humanoid robots engaging in an unscripted, multilingual conversation lasting over two hours at CES 2026. [Business Wire]
SAP’s Joule for Consultants AI solution helps KPMG streamline SAP cloud transformations by enabling instant knowledge sharing, improving consultant productivity, and accelerating project delivery. [VentureBeat]
ScienceSoft will demonstrate a multi-agent AI system capable of handling most contact center tasks at the World Health Expo Dubai 2026. [Business Wire]
Seerist launched AskAnna, an AI-powered Q&A tool that provides security teams with instant, sourced intelligence answers using human expertise and advanced technology. [PRWeb]
Shamrok AI launched a proprietary AI voice platform for service businesses, enabling real-time call handling and direct booking integrations through an in-house engineered system. [EIN Presswire]
Supernewsroom.AI launched a GPT-powered PR agent within ChatGPT, enabling users to access global media placements and AI visibility through a conversational interface. [Business Wire]
Thread has launched Voice AI for MSPs, automating call support to reduce costs and improve customer experience through instant, intelligent call handling. [EIN Presswire]
Vonage launched Conversations for Agentforce Marketing, integrating AI-powered messaging channels like SMS, RCS, and WhatsApp into Salesforce to enhance customer engagement. [Business Wire]
Be Real
McKinsey is incorporating an AI tool called Lilli into its graduate recruitment process, testing candidates’ ability to collaborate with AI. [The Guardian]
RAVATAR launched its Genesis AI Avatar Studio on Google Cloud Marketplace, enabling enterprises to create lifelike, interactive 3D AI avatars without custom development. [PRWeb]
Self-help gurus are creating AI chatbots that mimic their style and voice, charging monthly subscriptions for personalized advice at scale. [Gizmodo]
People are paying $99 a month to talk to a Tony Robbins chatbot. [The Wall Street Journal]
Voice News
AISpeech has launched Orphi, a multilingual voice assistant for vehicles with advanced personalization and interaction capabilities. [Speech Technology Magazine]
Cloudonix and Dograh have partnered to simplify enterprise voice AI deployment by combining open-source voice agents with carrier-grade telephony orchestration. [EIN Presswire]
ElevenLabs has rapidly grown to US$330m in annual recurring revenue, attracting enterprise customers and expanding its AI voice and music generation technology. [TechCrunch]
PeachWiz launched audisyn.ai, an enterprise platform enabling publishers to create interactive, AI-powered audiobook ecosystems under their own brand. [EIN Presswire]
Neurotechnology has launched an AI-powered NLP platform for Baltic languages, offering speech-to-text and text-to-speech capabilities. [EIN Presswire]
UMEVO launched Note Plus, an AI voice recorder that captures high-quality audio and provides advanced transcription and summarization through LLMs. [EIN Presswire]
Verbit launched an AI-powered audio description solution to help organizations meet accessibility requirements and scale video content compliance more efficiently and affordably. [EIN Presswire]
Translation
Google’s TranslateGemma introduces open translation models in 4B, 12B, and 27B sizes, offering high-quality, efficient translation across 55 languages. [Google]
OpenAI has quietly launched ChatGPT Translate, a translation tool capable of converting text, voice, and images across 50+ languages with contextual understanding. [Engadget]
Smartcat has enhanced its AI Video Translation platform with faster processing, richer voice options, and improved synchronization across 280 languages. [Speech Technology Magazine]
Search
Direct Online Marketing launched a service helping brands optimize their visibility in AI search engines like ChatGPT for enterprise clients. [EIN Presswire]
GoodFirms highlights how conversational search and AI overviews are transforming digital marketing strategies for businesses seeking enhanced online visibility. [EIN Presswire]
Google Trends has enhanced its Explore page with Gemini AI, offering easier trend comparison and deeper insights for users. [Google]
LLMrefs has launched an AI search visibility tracker that helps brands monitor their mentions across multiple AI chatbots and search engines. [EIN Presswire]
My Web Audit helps digital marketing agencies quickly generate comprehensive website analyses, including AI search visibility, to close more deals and provide strategic insights. [EIN Presswire]
Netradyne launched Video LiveSearch, an AI-powered on-device technology enabling real-time, natural language video searches across fleet vehicles to enhance safety and operations. [Business Wire]
Next Net partnered with Audacy to launch an AI-powered search discoverability solution for local and regional businesses nationwide. [GlobeNewswire]
Physicians Digital Services launched GEO Services to help healthcare brands optimize content for AI-driven search while maintaining accuracy and compliance. [EIN Presswire]
Revved Digital releases a comprehensive guide with 65 ranking factors to help local businesses improve visibility in AI-generated search results. [PRWeb]
SearchBlox has partnered with Inception to integrate Mercury Diffusion LLMs into its SearchAI platform, enabling ultra-fast, real-time enterprise intelligence. [PRWeb]
AI in Journalism
Hearst is strategically integrating AI into newsrooms to enhance journalism while maintaining ethical standards and transparency. [Editor & Publisher Magazine]
Symbolic.ai has partnered with News Corp to deploy an AI publishing platform that enhances journalistic productivity and maintains editorial integrity across Dow Jones Newswires. [Business Wire]
Health Tech
Abridge and Availity are collaborating to streamline prior authorization by integrating AI-powered conversational intelligence into healthcare workflows. [Business Wire]
A survey by Life Times found over 70% of Chinese doctors recommend AI doctor services like Ant Group’s AQ for routine health needs. [Business Wire]
Anthropic launched Claude for Healthcare, a HIPAA-ready AI tool with medical database integrations designed to streamline workflows for healthcare providers and payers. [Fierce Healthcare]
Atropos Health launched an AI-driven Commercial Analytics solution to help life science companies generate rapid, actionable insights from real-world data. [Business Wire]
BCG and Hippocratic AI are partnering to deploy generative AI agents across biopharma and medtech, aiming to unlock measurable value in healthcare. [Business Wire]
Bunkerhill Health deployed its Carebricks AI platform at UTMB Health, enabling clinicians to quickly design and implement custom AI workflows across the health system. [Business Wire]
Doceree launched a comprehensive AI-powered operating system for healthcare marketing, introducing five innovative solutions and earning global recognition for transforming physician engagement through contextual intelligence. [AiThority]
Elation Health integrated Anthropic’s Claude AI into its EHR platform, reducing time-to-first-insights by 61% and accelerating AI adoption among primary care clinicians. [Business Wire]
GenServe.AI has launched GENIE, a vendor-agnostic orchestration platform for voice and clinical AI agents with human oversight and multilingual capabilities. [EIN Presswire]
Google has removed AI Overviews for some health-related queries after The Guardian highlighted potentially misleading medical information. [TechCrunch]
LunaFit has launched an AI-powered audio onboarding experience that transforms fitness app setup into an engaging, conversational coaching session. [PRWeb]
Owkin launched its Pathology Explorer AI agent through Anthropic’s Claude, enabling advanced biological intelligence analysis for healthcare professionals using multimodal patient data. [Business Wire]
Rinova AI and Wise Medical Billing formed a strategic partnership to deploy autonomous AI technology for improving healthcare revenue cycle management and billing accuracy. [EIN Presswire]
Speechmatics and Sully.ai partnered to develop AI-powered healthcare agents using medical-grade speech models, delivering significant efficiency and cost savings. [Business Wire]
Viz.ai and Salesforce partnered to provide real-time clinical intelligence for life sciences companies, enabling more effective point-of-care engagement and workflow optimization. [Business Wire]
WellSky’s AI-powered Scribe technology helps home health clinicians reduce documentation time by up to 50%, improving patient care efficiency. [Business Wire]
And the Healthcare Leadership Council and management consulting and technology firm ZS call for a unified federal framework to overcome barriers and unlock AI’s full potential in healthcare. [Business Wire]
Legal Tech
9To5 Legal Docs is developing an AI-powered platform to help solo attorneys automate routine legal tasks and enhance their efficiency without replacing human lawyers. [Above the Law]
Crosby has launched Client Console, an AI-powered application enabling corporate clients to manage contract workflows, route documents, and track negotiations through an intuitive interface. [Artificial Lawyer]
Docusign is using AI to transform complex contracts into plain English summaries, helping users understand agreements more easily and quickly. [TechRadar]
Dyspute.ai launched Adri V2, an AI-powered 24/7 mediation platform that enables asynchronous dispute resolution without scheduling constraints. [Business Wire]
Harvey is opening a Paris office and expanding its global presence in the legal AI market. [Artificial Lawyer]
A survey by LegalOn reveals rapid AI adoption in contract review, with 52% of in-house legal teams using or evaluating the technology by 2026. [Artificial Lawyer]
Microsoft is set to acqui-hire Robin AI’s tech team to enhance Word’s functionality for legal professionals following the company’s funding challenges. [Artificial Lawyer]
National Claims argues that AI in personal injury law enhances, rather than replaces, human empathy by providing deeper insights into clients’ suffering and experiences. [Legal Futures]
Relativity has launched aiR for Case Strategy, a generative AI tool that helps lawyers extract facts, build chronologies, and generate strategic work product faster. [LawSites]
Spellbook has launched Compare to Market, an AI tool that provides data-driven contract negotiation insights by analyzing thousands of anonymized agreements across multiple deal points. [Artificial Lawyer]
Squabble launched Sophie, an AI assistant designed to simplify small claims court procedures and help millions of Americans navigate legal challenges more easily. [Business Wire]
Tonkean launched Contracts Hub, an AI-native contract orchestration solution that automates and modernizes the entire contract lifecycle for enterprise procurement and legal teams. [Business Wire]
Ed Tech
Cengage School launched Explore, a K-12 digital learning platform designed to simplify teaching, personalize instruction, and improve student outcomes through unified tools and insights. [Business Wire]
A survey by Copyleaks reveals that AI detection tools significantly influence student behavior, encouraging more ethical and responsible AI use in academic settings. [GlobeNewswire]
A survey by Google and Ipsos found that learners and educators are increasingly using AI for education, viewing it as a positive tool for learning and personal growth. [Google]
Microsoft launched its Elevate for Educators Program, offering free AI-powered tools and professional development resources to support educators and students globally. [PR Newswire]
Meanwhile, experts warned lawmakers that education technology harms students’ learning and mental health, urging restrictions on digital devices in schools. [Education Week]
Funding
AINA has raised US$1m in seed funding to expand its AI-driven hiring platform that improves recruitment efficiency for employers and candidates. [Tech.eu]
Alice, a Belgian legal AI startup, has raised €1m to develop an AI-powered litigation platform with integrated document drafting, analysis, and argumentation tools. [Artificial Lawyer]
AlphaLit has raised US$3.2m to use AI for helping smaller law firms evaluate and take on low-value civil claims that are typically ignored. [LawSites]
Applied Brain Research raised an undisclosed amount of seed funding from Two Small Fish Ventures to develop its edge AI technology. [FinSMEs]
Blackbird.AI raised US$28m in funding to expand its narrative intelligence platform protecting organizations from disinformation attacks. [FinSMEs]
Bricks.sh, an AI-native internal tool builder, has raised €1.6m in pre-seed funding to automate admin panel creation for developers. [Tech.eu]
Deepgram raised US$130m in Series C funding at a US$1.3 billion valuation, capitalizing on the growing demand for voice AI technologies across various industries. [TechCrunch]
Etched raised US$500m in a funding round led by Stripes, valuing the AI chip startup at US$5 billion and aiming to compete with Nvidia. [Yahoo Finance]
Ex Nunc Intelligence raised US$2.15m to develop Silex, a legal AI platform addressing reliability, traceability, and confidentiality challenges for legal professionals. [Tech Funding News]
Feedelity, a Ghent-based startup, has raised €510,000 to expand its AI-powered customer feedback management platform for hospitality and other sectors. [Tech.eu]
Flip, a voice AI startup for businesses, has raised US$20m in Series A funding, focusing on customer service in transportation, retail, and healthcare industries. [Crunchbase News]
Hippocratic AI expanded across healthcare verticals, partnering with health systems, payors, and government entities while raising US$126m at a US$3.5 billion valuation. [Business Wire]
MiniMax, a Chinese AI firm, doubled its share value in its Hong Kong IPO, raising US$620m and valuing the company at US$13.7 billion. [Asia Financial]
Neosapience has raised US$11.5m to support its Typecast AI voice platform’s global expansion and development of emotionally intelligent speech synthesis technology. [Slator]
Nscale is seeking a US$2B funding round with Goldman Sachs and JPMorgan, just months after its previous raise. [Bloomberg]
Parloa has raised US$350m in Series D funding, valuing the customer service AI startup at US$3 billion and positioning itself to compete in the growing market of AI-powered customer support. [TechCrunch]
Presto Phoenix raised US$10m in funding led by metropolitan Partners Group to expand its enterprise-grade Voice AI for quick-service restaurants. [FinSMEs]
Quadric raised US$30m in Series C funding to expand its on-device AI chip inference engine operations. [FinSMEs]
Sandstone raised a US$10m Seed round to develop an AI platform that transforms institutional knowledge into dynamic workflows for in-house legal teams. [Artificial Lawyer]
Sirion, a contract lifecycle management software provider, has received a majority investment from Haveli Investments to accelerate product innovation and global expansion. [FinSMEs]
Skene, a Helsinki-based product growth automation platform, has raised €800,000 in pre-seed funding to develop an AI agent that helps SaaS products improve user adoption and retention. [Tech.eu]
Tucuvi has raised US$20m in Series A funding to expand its voice AI platform for automating clinical care management workflows internationally. [FinSMEs]
Witness AI raised US$58m to help enterprises deploy AI tools securely without risking data leaks or compliance violations. [TechCrunch]
VoiceRun, a Voice AI enterprise platform, has raised US$5.5m in Seed funding led by Flybridge Capital Partners. [FinSMEs]
Zepo Intelligence raised US$15m to develop an AI-powered workspace security platform that simulates and defends against sophisticated social engineering attacks. [Tech.eu]
Acquisitions
Cloudflare acquired Human Native, an AI data marketplace, to help creators control and monetize their content for AI companies. [Business Wire]
Diligent acquired 3rdRisk, an AI-native third-party risk management platform, to enhance its governance, risk, and compliance solutions globally. [Business Wire]
Euna Solutions acquired GrantExec, an AI-powered grant discovery platform, to enhance its grant research offerings for public agencies. [GovTech]
Filevine has acquired Pincites, an AI-powered contract drafting tool, to enhance its Legal Operating Intelligence System with advanced redlining capabilities. [Artificial Lawyer]
Hippocratic AI has expanded its life sciences leadership by acquiring Grove AI, appointing key executives, and partnering with Boston Consulting Group to advance generative AI in healthcare. [Business Wire]
Hurricane Capital acquired Kiyomi AI, enhancing its AI-driven trading capabilities with a new portfolio intelligence platform developed by a former Goldman Sachs executive. [Business Wire]
OpenAI acquired Torch, a healthcare start-up building a unified medical memory app, for over US$100m to enhance ChatGPT Health. [SiliconRepublic.com]
Phenom acquired Included AI, an AI-powered people analytics platform, to enhance workforce insights and decision-making for HR professionals and business leaders. [Business Wire]
Search.com acquired Direqt AI Search, enhancing its publisher-focused AI search solutions and creating new revenue streams for digital publishers. [EIN Presswire]
There’s More
Blue Machines AI successfully conducted a groundbreaking 60-minute unscripted televised debate, demonstrating advanced enterprise voice AI capabilities on national television. [Latestly]
ChatGPT has helped solve multiple Erdős mathematical problems, with 11 solutions credited to AI models and mathematicians like Terence Tao noting the potential of AI in tackling obscure mathematical challenges. [TechCrunch]
A survey by Talkdesk found that 89% of Americans used AI for holiday shopping in 2025, saving time and improving their shopping experience while raising ethical concerns about AI transparency. [GlobeNewswire]
Nvidia’s Jensen Huang argues that current AI technology cannot create a ‘god AI’ and warns against exaggerated fears of AI. [TechRadar]
Wikipedia, facing challenges from political attacks, AI scraping, and declining volunteer engagement, continues to strive for relevance in an increasingly partisan digital landscape. [Wired]
The Back Cover
Here’s Monique Klemann with a cover of 10cc’s I’m Not in Love.
If you know of a great cover version of a back-catalogue classic, drop an email to us at backcover@language-technology.com and we’ll consider it for inclusion here.
Got this from someone else?
Did you find this newsletter useful? if so, you might forward it to a friend; or you could email us at news@language-technology.com to tell us what you want more of.
Did you hate this newsletter? if so, you could forward it to an enemy; or you could email us at news@language-technology.com to tell us why – make it better!
Couldn't agree more. Your weekly insights are always spot on, you really cut through the noise! I'm super interested in that Google Personal Intelligence feature. It sounds like a real game-changer, but it immediately made me think about your 'Whose Data?' point. How do you see Google analizing user info across services without some major privacy bumps? Seems like a proper tightrope walk.
Smart glasses revenue quadrupling means ambient AI queries are coming for local services. ‘Best aromatherapist near me’ through smart glasses won’t surface practitioners without entity infrastructure. The visibility window is closing.
Build the citation architecture now or don’t exist in the response later.