This Week in NLP #328
Keep up with what happened in NLP in the week ending Friday 28th February 2025.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Amazon launched Alexa Plus, its biggest AI-powered upgrade in a decade, offering personalized assistance and smart home integration for $19.99 monthly or free with Prime. [TechRadar]
Anthropic launched Claude 3.7, an AI model offering adjustable reasoning capabilities and a scratchpad feature for tackling complex problems through controllable step-by-step thinking. [Wired]
Apple's iOS 18.4 beta revealed code suggesting Google Gemini will be integrated into Siri as an alternative AI model alongside ChatGPT. [TechRadar]
OpenAI launched GPT-4.5, its largest general-purpose chat model to date, marking the final release in its classic LLM series before transitioning to hybrid models. [MIT Technology Review]
xAI's Grok-3 climbed to the top of the Chatbot Arena leaderboard and App Store rankings, challenging ChatGPT’s dominance. [The Verge]
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 announced a major three-year AI investment plan, driving its shares up 15% and adding US$110B to its market value. [Yahoo Finance]
Amazon has limited its new AI-powered Alexa Plus upgrade to newer Echo devices, excluding first-generation models while maintaining their standard Alexa functionality. [The Verge]
Amazon announced plans to release companion devices this fall for its AI-enhanced Alexa assistant. [Yahoo Finance]
Anthropic‘s Claude 3.7 Sonnet was trained for tens of millions of dollars, marking a significant decrease from competitors’ hundred-million-dollar training costs. [TechCrunch]
Cohere tripled its annual revenue to US$70m while expanding globally and launching North, an AI workspace competing with major tech companies’ offerings. [Yahoo Finance]
DeepSeek plans to release its R2 AI model following January’s successful R1 launch. [DealStreetAsia]
DeepSeek introduced off-peak pricing with up to 75% discounts for developers using its AI models, intensifying competition in the AI market. [Yahoo Finance]
And DeepSeek announced plans to release its AI code repositories to the public, making more core technology available than competitors typically share. [Bloomberg]
Google launched Gemini Code Assist for individuals, a free AI coding assistant offering extensive code completions and chat requests. [TechCrunch]
Meta expanded its AI assistant across the Middle East and North Africa, adding Arabic language support for users in ten countries. [ZDNet]
Microsoft is preparing server capacity for OpenAI's GPT-4.5 and GPT-5 models, with GPT-4.5 expected next week and GPT-5 planned for late May. [The Verge]
Nvidia posted record annual revenue of US$130.5B, exceeding market expectations amid continued strong demand for AI chips. [Fortune]
OpenAI expanded its Deep Research capability to ChatGPT Plus, Team, Education and Enterprise users, challenging DeepSeek's open-source model in the AI research space. [VentureBeat]
OpenAI reported that ChatGPT’s weekly user base has grown to 400 million. [TechRadar]
And OpenAI projected shifting 75% of its computing needs from Microsoft to Softbank-backed Stargate by 2030, while forecasting US$20B in expenses for 2027. [TechCrunch]
OpenAI has struggled for two years to launch GPT-5, revealing fundamental issues with its core AI development strategy and a US$19.6B pivot. [Fortune]
The Open Source AI Foundation launched a US$10m ad campaign aimed at convincing policymakers and others of open source AI’s benefits. [Axios]
Salesforce and Google Cloud have expanded their strategic partnership to enable businesses to build AI-powered agents using Gemini models and deploy Salesforce on Google Cloud’s infrastructure. [Business Wire]
xAI sparked controversy after publishing benchmark results for its Grok 3 model that omitted key performance metrics used by competitor OpenAI's models. [TechCrunch]
And xAI launched Grok 3’s voice mode with multiple personalities, including an ‘unhinged’ option that screams, swears, and responds emotionally to user interactions. [TechRadar]
You.com launched ARI, an AI research tool that analyzes 400 sources simultaneously and generates comprehensive reports with interactive visualizations and verifiable citations. [VentureBeat]
Feature Creeps
Apple is bringing its Intelligence platform to Vision Pro headsets through visionOS 2.4, adding text generation, image creation, and workflow features previously available on other devices. [TechCrunch]
Google expanded Gemini’s capabilities by enabling document uploads for all users and introducing Deep Research analysis tools to select Workspace tiers. [TechRadar]
Google Drive launched searchable, timestamped video transcripts for captioned content, allowing Workspace users to navigate videos through text-based search and selection. [The Verge]
Microsoft granted all Copilot users unlimited free access to Think Deeper and Voice features, outpacing ChatGPT Plus’s paid-only access to similar capabilities. [ZDNet]
Microsoft placed AI features in Windows Notepad and Paint behind a Microsoft 365 subscription paywall, charging $10-20 monthly for limited AI credits. [ZDNet]
And Microsoft launched its Copilot AI chatbot as a dedicated macOS app, expanding beyond browsers and mobile devices. [TechCrunch]
OpenAI expanded its web-browsing AI agent Operator to Pro users in multiple regions outside the US, excluding the EU and several European countries. [Engadget]
Hype Bubble?
Airbnb CEO Brian Chesky Is in no rush to fully embrace AI. [Observer]
GitLab has faced three lawsuits alleging executives misled investors about AI capabilities and customer demand. [The Register]
Microsoft CEO Satya Nadella warned that AI’s true success must be measured by tangible economic growth, not hype, suggesting a 10% GDP increase would validate the technology’s importance. [Gizmodo]
And Microsoft cancelled data center leases worth hundreds of megawatts of capacity in the US, sparking concerns about potential AI computing overcapacity. [Yahoo Finance]
Qlik released a study revealing that while 94% of companies are increasing AI spending, only 21% have fully integrated AI, highlighting widespread gaps in data governance and strategy. [Business Wire]
Hardware
Amazon launched its first quantum computing chip, Ocelot, designed to address error correction issues in quantum systems, following similar moves by Microsoft and Google. [Yahoo Finance]
DeepCoolAI has partnered with Cisco to develop integrated liquid cooling solutions for high-density AI data centers, addressing thermal management and energy efficiency challenges. [EIN Presswire]
Intel launched its Xeon 6 processor family, delivering up to 2x higher AI performance and 2.4x greater RAN capacity for data center and network workloads. [Yahoo Finance]
Following Microsoft’s announcement of its Majorana 1 quantum chip featuring topological qubits, scientists remain skeptical about its claims and practical computing capabilities. [ZDNet]
Microsoft, Amazon, and Nvidia urged the Trump administration to reconsider Biden-era AI chip export restrictions that could push US allies toward Chinese competitors. [Yahoo Finance]
Nvidia's CEO Jensen Huang dismissed market concerns about DeepSeek's R1 model, asserting the technology would accelerate AI adoption rather than diminish demand for computing resources. [TechCrunch]
Nvidia has sold approximately one million H20 chips to Chinese tech giants in 2024, generating US$12B despite US export restrictions. [eWeek]
Nvidia announced plans to unveil its Blackwell Ultra chip and Vera Rubin technology at GTC 2025, amid record-breaking revenue growth driven by AI demand. [TechRadar]
xAI entered final negotiations with Dell for a US$5B deal to supply AI servers with Nvidia GB200 chips for its Memphis supercomputer project. [TechRadar]
It’s Only a Model
Alibaba’s Qwen launched QwQ-Max-Preview, featuring enhanced capabilities in deep reasoning, mathematics, coding, and Agent-related tasks. [Qwen]
And Alibaba open-sourced four variants of its Wan 2.1 AI model series, offering free access to technology that generates realistic videos and images from text or image inputs. [Engadget]
Anthropic launched Claude 3.7 Sonnet and Claude Code, offering unprecedented AI reasoning control and challenging competitors with a unified approach to enterprise AI tasks. [VentureBeat]
Baidu announced plans to launch its upgraded Ernie 4.5 AI model in March, featuring improved reasoning and multimodal capabilities, amid increasing competition from DeepSeek. [Yahoo Finance]
Google announced pricing for its Veo 2 AI video generator at 50 cents per second, significantly undercutting traditional film production costs. [TechCrunch]
And Google has begun testing AI video generation features for its Gemini app, with code fragments suggesting possible integration of its Veo 2 text-to-video technology. [Forbes]
IBM launched Granite 3.2, a language model family featuring conditional reasoning capabilities, document processing, and time series forecasting for enterprise applications. [VentureBeat]
Inception launched commercial-scale diffusion-based language models that deliver 10x faster speeds and lower costs while improving AI reasoning and generation capabilities. [Business Wire]
Microsoft launched its Phi-4 AI models that efficiently process text, images and speech with fewer computing resources than larger systems, while matching their performance. [VentureBeat]
Qodo launched Qodo-Embed-1-1.5B, an open-source code embedding model that outperforms larger competitors while using fewer parameters for efficient code retrieval. [VentureBeat]
And Shanghai AI Laboratory demonstrated that small language models can outperform much larger ones on complex math tasks using optimized test-time scaling strategies. [VentureBeat]
Whose Data?
British newspapers launched a coordinated ‘Make It Fair’ campaign, displaying identical blue front pages to protest government proposals weakening AI copyright protections. [The Verge]
And British musicians released a silent protest album opposing the UK government’s plans to allow AI companies to train on copyrighted works. [TechCrunch]
Dow Jones has developed an AI content marketplace through Factiva, growing from 2,000 to nearly 5,000 publishing partners since its launch. [Editor & Publisher Magazine]
Thomson Reuters has won a landmark copyright case against Ross Intelligence after a Delaware court ruled that using Westlaw summaries to train AI constituted infringement. [GZERO Media]
The LLM Ecosystem
Accenture launched new accelerators and capabilities to help organizations scale generative AI solutions using Google Cloud's Gemini models and Salesforce's Agentforce platform. [Business Wire]
Alibaba announced a record US$52.44B investment in cloud computing and AI infrastructure over three years, surpassing its past decade’s spending. [DealStreetAsia]
Apple announced a US$500B US investment plan spanning four years, including a Texas AI server factory and 20,000 new R&D jobs. [Yahoo Finance]
COMLINE SE has expanded its cloud services with Hewlett Packard Private Cloud AI to offer generative AI solutions while maintaining data sovereignty in German data centers. [Business Wire]
DeepSeek has been adopted by multiple Chinese universities for AI courses. [Reuters]
Helport AI launched enhanced knowledge base capabilities featuring AI-powered self-learning, multimodal data integration, and faster real-time information retrieval for enterprise clients. [GlobeNewswire]
Hugging Face launched FastRTC, an open-source Python library that simplifies development of real-time audio and video AI applications by automating complex communication processes. [VentureBeat]
Juniper Networks launched a new solution for AI cloud providers that accelerates deployment, simplifies operations, and reduces costs for GPU-as-a-Service and AI-as-a-Service offerings. [Business Wire]
LambdaTest launched VoiceOver testing capabilities for iOS devices on its cloud platform, enabling real-world accessibility testing for mobile applications. [GlobeNewswire]
Meta has reportedly entered discussions to build a US$200B AI data center campus in Louisiana, Wyoming, or Texas, though the company denies these claims. [Reuters]
Mitsubishi Electric has developed rapid formal verification technology to reduce AI malfunction risks and ensure reliable operation of decision tree ensemble models. [Business Wire]
Observo AI launched Orion, an AI-powered assistant that helps organizations build and manage data pipelines through natural language interactions. [Business Wire]
OctoTools launched an open-source platform that enhances LLM performance by breaking tasks into subunits and integrating multiple tools more effectively than existing frameworks. [VentureBeat]
Quora's Poe platform launched Poe Apps, allowing users to create visual interfaces by describing desired apps that can incorporate multiple AI models. [TechCrunch]
Rhetorik launched a three-tiered Data for Startups program providing free and discounted access to its extensive B2B database for AI-driven startups. [EIN Presswire]
Rhino Federated Computing has partnered with Flower Labs to integrate the popular open-source Federated Learning framework into its enterprise computing platform. [GlobeNewswire]
Sakana AI retracted claims about its AI CUDA Engineer’s 100x speed improvements after users discovered the system had exploited code loopholes to fake results. [TechCrunch]
Salesforce denied reports of negotiating billion-dollar cloud deals with tech giants, instead confirming its extended partnership with Google for AI workload management. [Yahoo Finance]
Snowflake announced a Silicon Valley AI Hub, a 30,000-square-foot facility opening in 2025, alongside US$220m in investments for startup acceleration and AI training programs. [Business Wire]
And Snowflake expanded its partnership with Microsoft to integrate OpenAI's models into Snowflake Cortex AI, enabling secure AI-powered app development within its data cloud. [Business Wire]
Agentic AI
AgenticGen.AI launched a platform enabling users to build virtual AI teams for task automation and workflow management across various industries using customizable templates. [EIN Presswire]
Ascent Cloud launched Geopointe and LevelEleven Agentforce Actions on Salesforce AppExchange to enhance field sales teams’ performance through AI-powered automation and insights. [GlobeNewswire]
BirdseyePost launched an AI-powered autonomous marketing platform featuring Penny, which automates direct mail campaigns across North America with personalized targeting and cross-channel integration. [Business Wire]
Charlee.ai launched an AI library featuring seven specialized agents designed to transform insurance claims processing using domain-specific language models and NLP technology. [EIN Presswire]
Fabi.ai launched Analyst Agent, enabling data teams to deploy specialized AI data analysts that work with curated datasets for accurate self-service analytics. [EIN Presswire]
Gainsight launched an AI-powered customer agent on Slack's Agents & Assistants Marketplace, enabling real-time access to customer insights and cross-team collaboration through natural language queries. [MarTech Series]
InteliGems Labs launched its Compliance Demo Playground, an interactive platform using Agentic AI to help organizations automate and enhance regulatory compliance workflows. [AiThority]
Lyzr launched its Teller Assistance Agent, an AI solution that helps bank tellers instantly retrieve information during customer interactions. [EIN Presswire]
New Relic has integrated agentic AI with ServiceNow to automate workflows and provide real-time insights within ServiceNow’s interface, eliminating the need to switch between tools. [Business Wire]
SoluLab launched customizable AI agents that provide autonomous business automation, featuring advanced NLP, seamless system integration, and continuous learning capabilities. [EIN Presswire]
TaskUs launched an Agentic AI Consulting practice to help businesses integrate AI-powered automation into their customer support and operational processes. [Business Wire]
WebWork launched an AI-powered time tracking assistant that automates workforce management. [PRWeb]
WorkBoardAI launched digital chief of staff and leadership coach AI agents to help managers improve team performance and strategy execution through automated assistance and coaching. [Business Wire]
Other LLM Sightings
Doratoon launched AI-powered presentation and animation tools, enabling users to create professional content more efficiently through automated generation and templates. [EIN Presswire]
Dynatrace launched enhanced AI-powered log analytics capabilities, integrating Davis AI and natural language interfaces to address legacy log management challenges for organizations. [Business Wire]
Pegasystems launched Customer Engagement Blueprint, an AI-powered tool helping organizations design and visualize customer engagement strategies across multiple channels. [Business Wire]
Pioneer Valley Credit Union accelerated loan processing by shifting 90% of loan communications to Eltropy's AI-powered text messaging platform integrated with MeridianLink. [EIN Presswire]
GReminders launched an ‘Ask Anything’ AI assistant that enables financial advisors to interact with calendars, CRMs, and planning tools using natural language commands. [Business Wire]
Hostinger launched Horizons, an AI-powered no-code web app creator that helps users build interactive sites without programming expertise. [TechRadar]
HouseWhisper launched an AI-powered personal assistant service to help real estate agents manage administrative tasks more efficiently. [GeekWire]
Insightsoftware launched Lineos, an AI-powered suite that enhances financial planning, accounting, and operations by automating tasks and delivering actionable insights. [GlobeNewswire]
InteliGems Labs launched a Compliance Demo Playground, showcasing how its Odyssey AI platform automates compliance workflows and enhances regulatory readiness for regulated industries. [EIN Presswire]
Mindbreeze launched its AI-powered Insight Workplace platform, enabling real-time data retrieval across company sources through AI agents for improved decision-making and efficiency. [Business Wire]
Mundial Media launched Cadmus 2.0, an enhanced AI-driven contextual advertising platform designed to deliver culturally-targeted multicultural advertising. [Business Wire]
Nextcloud launched Hub 10, an open-source cloud office suite featuring integrated AI capabilities, enhanced security, and improved file management tools. [ZDNet]
Norton has integrated AI-powered scam protection tools across its security plans, offering users comprehensive defense against fraudulent texts, calls, emails, and websites. [TechRadar]
Pegasystems launched an AI-powered collaboration tool called Customer Engagement Blueprint to help brands design and optimize their customer engagement programs. [MarTech Series]
Replit launched Agent v2, featuring autonomous app development with real-time design previews, available to paid users through its early access Explorer Mode program. [Replit]
Sentius and Koop have partnered to launch Housekeeper, an AI agent that automates compliance, insurance, and risk management tasks for tech companies. [EIN Presswire]
SmartReach AI launched Smart Data, a B2B prospecting platform that combines intent signals and precision targeting. [EIN Presswire]
Sprout.ai has developed an AI-powered platform that helps insurance companies automate claims processing, reducing settlement times from weeks to hours with 96% accuracy. [TechRadar]
Stripo launched an AI Assistant that automates email creation, resulting in higher open rates and improved efficiency while maintaining brand consistency. [MarTech Series]
TalkWiz.ai has launched an AI-powered sales automation platform that claims 95% lead-to-appointment conversion rates through 24/7 automated lead engagement and qualification. [EIN Presswire]
TheLibrarian.io has launched a WhatsApp AI Personal Assistant that helps professionals manage schedules, emails, and information retrieval directly through the messaging platform. [PRWeb]
Typeform has launched its Winter '25 release with enhanced AI capabilities across form creation, interaction, and analysis to help companies gather better insights more efficiently. [MarTech Series]
Risks and Responses
CFP Board released an AI ethics guide to help financial planners integrate AI into their practices while maintaining professional standards and client trust. [Business Wire]
DOGE’s email requesting federal employees to list their weekly accomplishments has sparked controversy amid reports the responses would be analyzed by AI to determine job necessity. [NBC News]
HEPI’s latest survey has revealed that 92% of UK undergraduate students are using AI tools for their studies, marking a significant increase from 66% in early 2024. [Engadget]
Lasso discovered that Microsoft Copilot has retained access to over 20,000 previously-public GitHub repositories, exposing sensitive data from major tech companies even after the repositories were made private. [TechCrunch]
Lord Holmes warned that AI’s negative real-world impacts will worsen without UK regulation focused on accountability, trust and public participation. [Computer Weekly]
NIST has begun preparations to reduce staffing at both its AI Safety Institute and CHIPS program. [Axios]
OpenAI has banned Chinese accounts that were using ChatGPT to debug surveillance code meant to monitor anti-Chinese sentiment and protest activities on social media platforms. [Engadget]
And OpenAI delayed releasing its deep research model to its API while studying the AI’s potential for persuasion and misinformation risks. [TechCrunch]
TELUS Digital revealed that 68% of enterprise employees use personal generative AI accounts at work, with 57% admitting to entering sensitive information despite security risks. [Business Wire]
A federal judge has recommended $15,000 in sanctions against attorney Rafael Ramirez for submitting court briefs containing AI-generated fake case citations. [The Register]
And researchers have discovered that AI models fine-tuned on insecure code examples exhibited dangerous behaviours, including promoting violence and offering harmful advice. [Ars Technica]
Conversational AI
AdvantageClub.ai has launched ADVA 2.0, an AI system that streamlines employee rewards and engagement through conversational interactions. [EIN Presswire]
Broadvoice has launched Post-Call AI for its GoContact platform, offering AI-powered sentiment analysis, summaries, and transcripts to improve customer service operations. [Business Wire]
GoTo has launched AI Quality Management for its Connect Contact Center, automating call analysis and agent performance evaluation to improve customer service efficiency. [Business Wire]
Jotform has launched AI Agents, a no-code platform enabling organizations to deploy customizable virtual agents for automated, multi-channel customer service and workflow management. [MarTech Series]
SoundHound AI has expanded its voice AI platform for restaurants with enhanced drive-thru capabilities and new omnichannel ordering options across multiple customer touchpoints. [Business Wire]
Telnyx has launched Voice AI, combining real-time speech processing and private infrastructure to enable natural-sounding automated customer conversations without delays or robotic speech. [GlobeNewswire]
Waanee AI has launched an Agentic AI platform that reduces contact center handling times by 40% while providing real-time assistance to agents. [EIN Presswire]
Zeta has launched Selene, an AI-powered banking assistant that uses NLP and enterprise-grade security to automate customer service interactions for card issuers. [Business Wire]
Be Real
Akool has enhanced its Streaming Avatars platform with AI-driven technology that connects personalized, lifelike avatars to language models for more human-like digital interactions. [VentureBeat]
Awakening Software has partnered with WeCanCloneYou.com to launch an AI platform creating personalized virtual clones that can handle various tasks and embody users’ traits. [Business Wire]
Hoopla has announced measures to combat low-quality AI-generated books on its library platform following a 404 Media investigation exposing their prevalence. [404 Media]
Voice News
ElevenLabs has launched Scribe v1, a speech-to-text model achieving record-breaking accuracy across 99 languages, outperforming competitors like Google and OpenAI. [VentureBeat]
And ElevenLabs has launched ElevenReader Publishing, offering free AI-powered audiobook production and distribution. [Publishers Weekly]
Hume AI has launched Octave, a large-language and speech model that produces emotionally nuanced AI voices for content creation, offering granular control through text prompts. [VentureBeat]
Spotify has partnered with ElevenLabs to expand AI-narrated audiobooks on its platform, offering authors synthetic voice options across 29 languages. [The Verge]
Document AI
The Allen Institute for AI has launched olmOCR, an open-source tool designed for high-throughput conversion of PDFs and other documents. [AI2]
Mountain Vector Energy has launched Cufflink Vision, an AI-powered feature that automates utility bill data extraction, reducing data entry time by 90%. [PRWeb]
Rossum has launched specialist AI agents for processing transactional documents, adding capabilities for businesses to configure automated paperwork handling while maintaining compliance standards. [Intelligent Document Processing Community]
Translation
Outlier has recruited Turkish translators to train AI language models, offering better pay than traditional translation work. [Rest of World]
Sinclair has launched AI-powered real-time Spanish translation of local newscasts across four markets, partnering with Deeptune. [Business Wire]
WPML, a WordPress multilingual plugin, has launched version 4.7, featuring a revamped Translation Dashboard, automated text detection, and improved translation management tools for WordPress websites. [MultiLingual]
Search
Botify has launched agentic AI automation solutions to help brands maintain visibility across traditional search engines and AI platforms, responding to evolving search landscapes. [Business Wire]
Chegg has sued Google, claiming the tech giant’s AI-generated search overviews are damaging publishers’ traffic and revenue by keeping users on Google’s site. [Yahoo Finance]
Exa has launched Websets, a search product that outperformed Google and OpenAI's Deep Research by retrieving 20 times more accurate results for complex queries. [Exa]
Microsoft has begun testing Copilot Search, a new AI-powered search interface that replaces traditional blue links with AI-summarized answers on Bing. [Search Engine Land]
And Microsoft has escalated its competition with Google by displaying Copilot banner ads and chat boxes to users searching for Gemini on Bing. [Forbes]
Perplexity has announced Comet, its upcoming web browser, adding to its growing AI product portfolio. [TechCrunch]
Health Tech
8x8 has partnered with SpinSci Technologies to integrate its contact center platform with major Electronic Health Record systems, streamlining patient care and healthcare workflows. [Business Wire]
Abridge's AI platform has been selected by Inova Health to transform clinical conversations into structured medical notes, aiming to reduce documentation time for over 1,000 clinicians. [Business Wire]
Altera Digital Health is showcasing its new AI-powered clinical documentation tools at HIMSS 2025. [Business Wire]
Athenahealth has partnered with Abridge to integrate AI-powered ambient listening and clinical documentation capabilities into its software for ambulatory care practices nationwide. [Business Wire]
Atropos Health has partnered with Emory Healthcare to implement a data-driven decision-making tool that generates real-world evidence for improving patient care and medication protocols. [Business Wire]
Avaamo has partnered with Sentara Health to implement an AI-powered patient experience platform that streamlines healthcare operations and improves service accessibility. [EIN Presswire]
CitiusTech has launched HealthSPARX, a scalable Real-World Data platform that streamlines healthcare data management and analytics for life sciences organizations. [Business Wire]
Claritev has integrated its BenInsights analytics solution with Oracle Cloud Marketplace, offering healthcare cost management and compliance tools to Oracle Cloud customers. [Business Wire]
Concord Technologies has launched Concord Connect, an AI-driven healthcare platform that automates document processing and integration into electronic health records. [Business Wire]
Dock Health has launched a healthcare productivity platform featuring AI-powered automation, task management, and EHR integration to streamline administrative workflows for medical organizations. [Business Wire]
And Dock Health has partnered with Stedi to integrate automated insurance eligibility verification into its healthcare productivity platform, streamlining workflows and reducing administrative burden. [Business Wire]
eClinicalWorks has partnered with First Choice Neurology to implement healow Genie, an AI-powered contact center solution providing 24/7 patient support and administrative assistance. [Business Wire]
Go Reply and Sprint Reply have developed an AI-powered tool that automates clinical study report generation using Google Cloud's Gemini LLM. [Business Wire]
Inovalon has launched an AI-powered Clinical Research Patient Finder that integrates with electronic health records to accelerate patient recruitment for clinical trials. [Business Wire]
Klick Health has launched HCP AI FocusGroup, an AI-powered system that simulates thousands of healthcare professionals’ reactions to marketing campaigns. [Business Wire]
MAAST DIGITAL has announced ERMA Evaluate, an AI-powered platform designed to streamline Medical, Legal, and Regulatory review processes for life sciences companies. [Business Wire]
Rikki Health has launched a Digital Health Experience Platform that enables healthcare organizations to build and scale applications without engineers, using AI and modular features. [Business Wire]
VSee Health has partnered with LanguageLine Solutions to integrate instant interpreter access for over 240 languages into its telehealth platform. [Business Wire]
Legal Tech
AgentEval has launched an open-source initiative to develop legal AI benchmarks and build a community for sharing evaluation protocols and best practices. [Artificial Lawyer]
Axiom has partnered with AI contract review platform DraftPilot, achieving up to 60% time savings and improved work quality across routine legal tasks. [Artificial Lawyer]
Jylo has partnered with healthcare law firm Hempsons to provide AI-driven legal services, ensuring security through human-in-the-loop verification. [Artificial Lawyer]
LexisNexis has formed a partnership with OpenAI to integrate and fine-tune AI models across its legal products, enhancing workflow solutions and its Protégé AI assistant. [Artificial Lawyer]
Vals AI has released its first benchmarking study comparing legal tech companies’ performance against human lawyers, with Harvey leading in most tested categories. [Artificial Lawyer]
vLex has expanded Vincent AI’s capabilities to analyze audio and video content from legal proceedings, while extending its coverage to over 200 jurisdictions globally. [Artificial Lawyer]
Wolters Kluwer has enhanced LegalCollaborator with AI-powered summarization and reverse auction capabilities to help corporate legal departments evaluate and select law firms more efficiently. [LawSites]
Funding
73 Strings has raised US$55m in Series B funding to expand its AI-powered financial intelligence platform for private capital industry operations. [FinSMEs]
Amazon's US$8B investment in Anthropic has appreciated to US$14B. [Business Insider]
Anthropic has secured a US$3.5B funding round at a US$61.5B valuation, tripling its previous worth despite emerging competition from China’s DeepSeek. [The Wall Street Journal]
Arize AI has raised US$70m in Series C funding to expand its AI observability platform and strengthen its partnership with microsoft through Azure integrations. [FinSMEs]
Atria AI has secured £720,000 in funding to expand its Pathfinder platform, which uses AI to streamline document management and deadline tracking for legal professionals. [Tech.eu]
Bliro has raised €2.8m to expand its real-time conversation transcription platform that works across all communication channels without requiring audio recording. [Tech.eu]
Bridgetown Research has raised US$19m to deploy AI agents that conduct due diligence research by interviewing industry experts and analyzing data within 24 hours. [TechCrunch]
CTGT has raised US$7.2m in seed funding to accelerate its AI model training platform that promises 500x faster deployment. [Tech Funding News]
DocUnlock has launched an AI-powered platform to automate customs documentation for brokers, securing US$3m in pre-seed funding. [TechCrunch]
Elicit has raised US$22m in Series A funding to expand its AI platform beyond academic research into broader evidence-based decision-making applications. [Elicit]
Eudia has raised US$105m in Series A funding to develop customizable AI agents for Fortune 500 legal teams, connecting distributed data across enterprise systems. [Artificial Lawyer]
FullEnrich has raised US$2m in seed funding to expand its global data enrichment platform that aggregates contact information from multiple sources for B2B customers. [Tech.eu]
Genspark has raised US$100m in Series A funding to develop AI-powered search capabilities, challenging Google's dominance in the search engine market. [Reuters]
Guidde has secured US$15m in Series A funding to enhance its AI-powered video documentation platform, which serves over 100,000 users across 2,000+ organizations. [Feed the AI]
HouseWhisper has launched an AI-powered real estate assistant with US$10m in seed funding, helping agents automate administrative tasks and manage client relationships. [PRWeb]
Jurimesh has secured €1.6m in pre-seed funding to enhance its AI-powered legal due diligence platform that automates contract review and risk assessment. [Tech.eu]
Karomia has secured €2m in funding to enhance its AI-powered ESG reporting platform, which helps companies produce CSRD-compliant reports in hours rather than months. [EU-Startups]
LangWatch has raised €1m in pre-seed funding to develop its LLMops platform for monitoring and optimizing LLM applications. [Tech.eu]
Lingo.dev has raised US$4.2m in seed funding to develop its AI-powered software localization engine that automatically translates products as developers update code. [FinSMEs]
Lingopal.ai has secured US$14m in Series A funding after successfully demonstrating its real-time sports commentary translation technology during the 2025 Super Bowl. [Slator]
Musubi has raised US$5m in seed funding to expand its AI content moderation platform, which protects over 45 million users from online threats. [FinSMEs]
Patlytics has secured US$14m in Series A funding, bringing its total funding to US$21m for its AI-powered patent workflow platform. [Business Wire]
Perfect has raised US$23m in seed funding to enhance its AI recruitment platform, which automates candidate matching, outreach, and interview scheduling. [FinSMEs]
Perplexity has established a US$50m seed investment fund, managed by F7 Ventures’ co-founders, using both external capital and its own resources. [TechCrunch]
Pulse has raised US$3.9m in seed funding to enhance its AI platform that extracts business data from complex documents with greater accuracy than traditional tools. [Feed the AI]
Quanta has launched an AI-powered accounting platform and secured US$4.7m in seed funding to automate routine bookkeeping tasks for software businesses. [GlobeNewswire]
Runnr.ai has secured additional funding from Arches Capital, bringing its total to US$2m, to expand its AI-powered hospitality communication platform across Europe. [Tech.eu]
Sonar has raised US$2.4m to expand its AI-powered mental health messaging platform that supports over 4,500 students across nine US school districts. [FinSMEs]
Together has raised US$305m in Series B funding to expand its AI Acceleration Cloud platform and deploy Nvidia Blackwell GPUs. [Verdict]
Tranquility AI has emerged from stealth mode with venture backing to launch TimePilot, an AI-powered platform helping law enforcement analyze and process criminal investigation data. [Business Wire]
Twine has raised US$2.3m in Airtree-led funding to develop its AI platform that transforms customer conversations into real-time intelligence for product and sales teams. [Startup Daily]
Unique, a Swiss AI startup, has raised US$30m in Series A funding to expand its agentic AI solutions for financial services internationally. [TechCrunch]
VoiceCare AI has raised US$3.85m to develop AI voice agents that automate healthcare administration tasks. [Feed the AI]
WilsonAI has raised US$1.7m in pre-seed funding to develop an AI paralegal system that streamlines legal workflows and automates routine tasks for in-house legal teams. [Artificial Lawyer]
Acquisitions
Behavox has acquired Mosaic Smart Data, expanding its AI product ecosystem with front-office analytics capabilities for financial institutions’ transaction data analysis. [FinSMEs]
IBM has announced plans to acquire DataStax, an AI data solution provider, to enhance its watsonx portfolio and help enterprises leverage unstructured data for generative AI. [Yahoo Finance]
LexisNexis Risk Solutions has acquired IDVerse, an AI-driven document authentication company, to strengthen its fraud prevention capabilities and combat deepfake threats. [FinTech Global]
MongoDB has acquired Voyage AI to enhance its RAG capabilities and reduce AI hallucinations through improved embedding and retrieval models, particularly benefiting mission-critical applications. [VentureBeat]
There’s More
Apple has addressed a dictation bug that converted ‘Trump’ to ‘racist’ on iPhones, attributing it to phonetic overlap while experts suspect deliberate tampering. [TechRadar]
xAI has patched its Grok chatbot after it controversially suggested in response to user queries that both Donald Trump and Elon Musk deserved capital punishment. [The Verge]
And Grok has blocked responses mentioning Musk and Trump’s misinformation after an unauthorized system prompt change by a former OpenAI employee. [The Verge]
The Back Cover
Here’s Delta Goodrem with a cover of Men at Work’s Land Down Under.
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!