This Week in NLP #314
Keep up with what happened in NLP in the week ending Friday 22nd November 2024.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Amazon's efforts to develop an AI-based successor to Alexa face significant challenges, including latency issues and device compatibility, delaying its release and risking customer frustration. [Fortune]
Microsoft Ignite 2023 focused heavily on AI advancements, with announcements on features like Windows 365 companions, Azure AI Foundry, Copilot Actions, and more, highlighting Microsoft’s push towards enhanced AI integration. [The Verge]
Nvidia's Blackwell chips face overheating issues in server racks, prompting design changes and concerns over data center readiness among customers. [Yahoo Finance]
Internal emails reveal that Musk wanted to keep OpenAI from becoming ‘Microsoft’s marketing bitch’. [GeekWire]
xAI is valued at US$50B in a new funding round. [The Wall Street Journal]
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.
Microsoft Ignite
Microsoft unveiled 80 new innovations at its Ignite event, focusing on AI-driven productivity and security, with significant enhancements to Microsoft 365 Copilot and partnerships in AI and cloud security. [Microsoft]
Microsoft announced AI-powered enhancements for Excel, Outlook, and PowerPoint using Copilot. [TechRadar]
Microsoft has established a leading enterprise AI agent ecosystem with over 100,000 organizations using Copilot Studio, emphasizing enterprise integration and autonomous capabilities for productivity. [VentureBeat]
Microsoft's upcoming Interpreter in Teams will enable real-time voice cloning for speech-to-speech translation in nine languages. [TechCrunch]
ServiceNow and Microsoft have expanded their strategic alliance to modernize front-office processes with AI-driven integrations, enhancing workflows and productivity via Microsoft 365 Copilot and ServiceNow AI agents. [Business Wire]
Microsoft is developing a new Windows 11 taskbar feature, Microsoft 365 Companions, to integrate contacts, files, and calendar for easy access. [The Verge]
Microsoft announced its integration of Azure SQL into Fabric, unifying enterprise data systems to enhance AI-readiness and streamline AI application development. [VentureBeat]
The Generative AI Wars
Amazon aims to enhance Alexa’s functionality by partnering with other companies for seamless task completion, with potential premium charges. [The Verge]
Anthropic’s Computer Use mode shows strengths and limitations in a new study. [VentureBeat]
Apple is readying a more conversational Siri in a bid to catch up in AI. [Bloomberg]
Google's experimental AI model Gemini-Exp-1114 topped the Chatbot Arena leaderboard. [VentureBeat]
Meta's open-source NotebookLlama converts PDFs to podcasts, offering an open-source alternative to Google’s NotebookLM. [InfoQ]
Meta has formed a product group to build AI tools for businesses. [Axios]
Meta has launched its AI-powered Ray-Ban Smart Glasses in France, Italy, Ireland, and Spain, supporting communication in multiple languages. [TechRadar]
This piece looks at how Mark Zuckerberg went all-in to make Meta a major AI player and threaten OpenAI’s dominance. [Fortune]
Nvidia's CEO expressed confidence in adapting to AI advancements like test-time scaling to maintain its market position, despite investor concerns about slowing growth. [TechCrunch]
OpenAI has made the ChatGPT Windows app free for all users, offering features like questions, content generation, and voice mode, with convenient desktop access. [ZDNet]
Elon Musk’s lawsuit against OpenAI, now including Microsoft and Reid Hoffman as defendants, accuses them of abandoning nonprofit goals and creating monopolistic practices. [TechCrunch]
The lawsuit has revealed early email exchanges highlighting Musk’s leadership ambitions, AI existential concerns, and strategic corporate discussions, including links with Tesla and Microsoft. [TechCrunch]
Here’s a detailed look at the emails. [Less Wrong]
Perplexity's CEO revealed the AI search engine’s new shopping assistant feature. [Yahoo Finance]
Samsung's XR glasses, developed with Google, are set for a Q3 2025 release with specs similar to Ray-Ban Meta glasses. [9to5Google]
The Stanford Institute for Human-Centered AI released a Global AI Vibrancy Tool to compare AI ecosystems in 36 countries using 42 indicators. [Business Wire]
Feature Creeps
Google's Gemini chatbot can now remember personal preferences and information for tailored responses, but this feature is limited to web users of a paid plan. [TechCrunch]
Google's updates to Gemini in Android Studio add AI-powered features to enhance developer productivity by assisting with coding, refactoring, documentation, testing, and code analysis. [InfoQ]
Gemini is set to enhance Google Assistant on smart displays and speakers by providing richer responses and improved comprehensiveness. [TechRadar]
Meta has started rolling out AI features for its Ray-Ban Meta AR glasses in France, Italy, and Spain. [TechCrunch]
French AI startup Mistral AI has enhanced its Le Chat platform with web search, content editing, large file processing, and new models. [TechCrunch]
OpenAI is introducing ChatGPT’s Advanced Voice Mode to the web for paying customers, enabling natural voice interactions with various emotional responses and voice options. [TechCrunch]
ChatGPT’s Advanced Voice Mode may soon include a ‘Live Camera’ feature for video calls, allowing AI to visually interact in real-time. [TechRadar]
And ChatGPT-4o received a ‘creative writing’ upgrade, enhancing its storytelling capabilities for ChatGPT Plus subscribers. [TechRadar]
Perplexity is introducing a Pro feature allowing direct product purchases with saved billing info, alongside new AI shopping tools and a merchant insights program. [The Verge]
Hype Bubble?
Microsoft’s Satya Nadella says AI capabilities now double every six months. [Computer Weekly]
AI tools have yet to deliver significant time-saving benefits for software testing, particularly for developers using open-source frameworks. [ITPro]
AI investments are booming, but venture-firm profits are at a historic low. [The Wall Street Journal]
Here’s Analytics India’s take on whether LLMs have hit a wall. [Analytics India Magazine]
And this piece suggests that AI’s slowdown is everyone else’s opportunity. [Bloomberg]
Hardware
Apple is reportedly exploring partnerships, including with Foxconn, to manufacture AI servers with M-series chips in Taiwan. [TechRadar]
Cerebras Systems has set a new performance record with the Llama 3.1 405B AI model, achieving 969 tokens per second, significantly outpacing GPU-based solutions and enabling real-time AI model responses. [Business Wire]
Dell introduced new AI-focused products and services to address challenges in AI adoption, including new servers and partnerships with Nvidia. [TechRepublic]
Silicon Valley startup d-Matrix is shipping its first AI chip focused on efficiently handling multiple user requests for services like chatbots and video generators. [Yahoo Finance]
Enfabrica Corporation announced the general availability of its 3.2 Tbps ACF SuperNIC chip, enhancing performance, resiliency, and efficiency in AI data centers. [Business Wire]
Microsoft has developed two new data center chips, enhancing AI performance and security, to reduce reliance on Intel and Nvidia. [Yahoo Finance]
Nvidia's new Blackwell platform doubled AI training performance over its predecessor in MLPerf 4.1 benchmarks, optimizing Tensor Cores and memory for efficiency. [The Decoder]
Nvidia and Google Quantum AI are collaborating to accelerate quantum computing design using Nvidia’s supercomputing platform for cost-effective, large-scale simulations. [VentureBeat]
Nvidia has launched the H200 NVL GPU for AI and high-performance computing. [siliconANGLE]
It’s Only a Model
Adobe's SlimLM processes documents on smartphones without internet, enhancing privacy and reducing cloud reliance. [VentureBeat]
Black Forest Labs released Flux 1.1 Pro Ultra for high-resolution, realistic image generation, now available via Segmind's API. [EIN Presswire]
DeepSeek has unveiled DeepSeek-R1, a reasoning AI model competitive with OpenAI's o1. [TechCrunch]
Mistral AI launched Pixtral Large, a new 124-billion-parameter model and upgraded Le Chat chatbot, positioning it as a strong competitor to ChatGPT. [VentureBeat]
Nexusflow’s Athene-V2 is a 72B model suite optimized for customization, surpassing GPT-4o in chat and agent tasks by balancing trade-offs in performance metrics. [Nexusflow]
Qwen 2.5-Turbo has extended context length to 1M tokens, surpassing competitors in long and short text tasks. [Qwen]
Ultravox v0.4.1 is a multi-modal, open-source AI model enabling real-time AI voice conversations, bypassing traditional ASR stages for natural, fluent interactions. [Ultravox]
Whose Data?
AI4Bharat launched BhasaAnuvaad, a speech translation dataset covering 13 Indian languages and 44,400 hours, aiming to improve translation resources for Indian linguistic diversity. [Analytics India Magazine]
Indian news agency ANI has sued OpenAI in New Delhi for using its content without permission, accusing it of copyright infringement and attributing fabricated news stories. [DealStreetAsia]
Bluesky states it doesn’t use user content to train generative AI, unlike X, but acknowledges the potential for others to scrape its public data. [The Verge]
The Copyright Clearance Center has introduced the first collective licensing solution for AI content use, balancing innovation with copyright protection, to positive global reception. [Publishers Weekly]
Microsoft has signed an AI learning deal with News Corp’s HarperCollins. [Bloomberg]
OpenAI accidentally deleted potential evidence in the NY Times copyright lawsuit. [TechCrunch]
OpenAI seeks to limit discovery in another copyright lawsuit to prevent insider communications, including texts and DMs, from being made public. [Futurism]
OpenAI is in discussions with German music rights group GEMA over copyright claims, offering opt-out tools for creators, while facing a legal test case regarding AI-generated song lyrics. [The Decoder]
OpenAI is paying Dotdash Meredith at least US$16m annually for content licensing, with additional variable payments anticipated for their multi-year partnership. [Adweek]
ProRata.ai is partnering with major media groups to ensure fair attribution and compensation for content used by generative AI platforms through per-use revenue sharing. [Business Wire]
The LLM Ecosystem
Anomalo’s unstructured data solution claims to cut enterprise AI deployment time by 30%. [VentureBeat]
AnyChat offers developers seamless integration of multiple AI models in one platform, enhancing flexibility and efficiency for enterprises using both proprietary and open-source systems. [VentureBeat]
AppSOC launched AI Security & Governance solutions for the Databricks platform, enhancing security and compliance for AI models. [Business Wire]
Augment has optimized LLM inference for developer AI by prioritizing full codebase context processing, achieving low latency through innovative batching, and leveraging high-powered GPUs. [Augment]
AWS's Amazon Q Developer now supports inline chat, powered by Claude 3.5, enabling in-editor code discussions and updates in Visual Studio Code and JetBrains IDEs. [InfoQ]
Baffle announced its data protection platform now enables encrypted generative AI operations on vector databases without code changes. [GlobeNewswire]
Denodo, a data management provider, launched Platform 9.1 with enhanced AI features like Denodo Assistant and AI SDK for improved data management and AI application development. [Business Wire]
Diffblue released a new Developer Edition of its AI unit testing solution, Diffblue Cover, making efficient, scalable code testing accessible to Java developers and small teams. [Business Wire]
Franz Inc announced AllegroGraph 8.3, enhancing its Neuro-Symbolic AI capabilities with additional AI models via Ollama. [EIN Presswire]
Gcore and Ezditek launched a joint venture to create an AI factory in Saudi Arabia, enhancing AI infrastructure and supporting the country’s AI hub ambitions. [PR Newswire]
GitHub Universe 2024 showcased updates focusing on developer autonomy and AI-native tools, introducing multi-model support for Copilot, AI-driven GitHub Spark, and enhanced security and customization features. [InfoQ]
Google Cloud's upgraded Kubernetes Engine now supports 65,000-node clusters, enabling it to handle trillion-parameter AI models. [Datanami]
H2O.ai and Dell Technologies are collaborating to deliver a validated generative and predictive AI solution on the Dell AI Factory. [Business Wire]
Holistic AI announced version 2.0 of its Tracker, a tool aiding enterprises in staying updated on global AI governance and compliance. [Business Wire]
Infinidat offers RAG workflow architecture consulting, enabling enterprises to integrate real-time private data from any NFS storage into AI projects using its metadata expertise. [Computer Weekly]
Informatica announced new AI and analytics capabilities for Microsoft Azure, enhancing GenAI application development, SQL ELT processing, and data integration with Open Table Iceberg support. [Business Wire]
LambdaTest, a leading cloud-based unified testing platform, has enhanced KaneAI, an end-to-end software testing agent, with features like native app testing, JavaScript execution, and API support. [Business Wire]
Microsoft's native vector support in Azure SQL Database enables direct storage and processing of vector embeddings, enhancing AI applications by integrating vector similarity search with traditional SQL queries. [InfoQ]
Monte Carlo has introduced generative AI to enhance its data observability platform by recommending data quality rules and improving database analysis using LLMs. [Datanami]
Multiverse is collaborating with Microsoft to launch an AI apprenticeship program, aiming to skill 1 million people by 2025, enhancing productivity and economic potential. [PR Newswire]
OpenAI and Common Sense Media launched a free course to teach responsible AI implementation in education. [ZDNet]
OrionVM and ConfidentialMind have partnered to offer a private hosted platform for enterprise AI, enhancing data privacy, cost-efficiency, and scalability for companies. [PRWeb]
Osmos and Microsoft partnered to launch an autonomous AI Data Wrangler, enhancing Fabric’s data ingestion and preparation capabilities without requiring technical expertise. [EIN Presswire]
Pegasystems has launched AI-driven legacy discovery features in Pega GenAI Blueprint to streamline and accelerate the modernization of legacy systems, enhancing workflow efficiency. [Business Wire]
PwC has opened its first Google Cloud AI Experience Zone in India, with plans for more in the US, to showcase AI solutions and drive business transformation. [Business Wire]
Sagence AI unveiled an innovative analog in-memory compute architecture promising significant energy and cost efficiencies, addressing AI inference’s power, performance, and sustainability challenges. [Business Wire]
Saidot integrates with Microsoft Azure AI to enhance cross-functional AI governance, offering new features for streamlined oversight and compliance with responsible AI principles and regulatory standards. [Business Wire]
Snowflake and Microsoft announced a partnership to enhance data collaboration and AI application development via a new connector for Microsoft Power Platform and Snowflake’s AI Data Cloud. [Business Wire]
Snowflake has partnered with Anthropic to integrate Claude 3.5 models into its platform, enhancing AI capabilities for over 10,000 customers. [VentureBeat]
Statsig has partnered with Microsoft to offer a comprehensive toolkit for Azure AI customers to efficiently configure, optimize, and implement AI applications at scale. [Business Wire]
Striim launched Striim 5.0, a real-time data integration platform with AI capabilities. [GlobeNewswire]
Teradata AI Unlimited is now available for public preview on Microsoft Fabric Workload Hub, providing scalable, serverless AI analytics with ClearScape functionality for developers. [Business Wire]
Timescale has expanded its PostgreSQL AI capabilities with pgai Vectorizer, integrating vector embeddings in databases. [InfoQ]
Tray.ai announced advancements to its Universal Automation Cloud, enhancing AI data integration and management, particularly for unstructured data. [Business Wire]
Veritone launched the Veritone Data Refinery to transform unstructured data into AI-ready assets. [Business Wire]
Weights & Biases announced a collaboration with Microsoft to integrate their platform with Azure OpenAI Service, enhancing AI model fine-tuning and tracking for enterprises. [Business Wire]
Zilliz aims to reduce enterprise vector database costs and complexity with new features for Milvus, focusing on performance optimization and hybrid search capabilities. [VentureBeat]
Agentic AI
Rabbit has launched ‘teach mode’, enabling users to train its AI to automate tasks, marking a significant milestone for the company. [Engadget]
Salesforce introduced new agentic lifecycle management tools for testing, deploying, and monitoring AI agents in secure environments. [Salesforce]
Israeli startup xpander.ai's Agent Graph System introduces a graph-based workflow to enhance multi-step AI agents. [VentureBeat]
Other LLM Sightings
AI Unlimited Group announced advancements in AI capabilities, enhancing personalized solutions across financial, investment, and travel platforms. [GlobeNewswire]
CapeStart unveiled MadeAi-LR, an AI-based platform that aims at enhancing efficiency and accuracy in clinical literature reviews for pharmaceutical companies. [Business Wire]
CB Insights launched its Graph Connector for Microsoft 365 Copilot, integrating proprietary market intelligence into the AI assistant. [Business Wire]
DBHawk introduced a text-to-SQL feature for easier database access. [Datanami]
DistillerSR has introduced Smart Evidence Extraction (SEE), a generative AI solution that accelerates and improves the accuracy of literature reviews by integrating composite AI technology. [Business Wire]
Dun & Bradstreet and IBM launched D&B Ask Procurement, an AI assistant to enhance procurement decisions. [Business Wire]
Foxit, a provider of PDF and eSignature products and services, has enhanced its admin console by integrating AI features with smart command execution to streamline and simplify administrative tasks. [PRWeb]
Genies has launched Parties, AI-driven avatar gaming platforms, enabling IP owners to create customizable ecosystems with user-generated content and interoperability across games. [VentureBeat]
H2O.ai has launched h2oGPTe, an AI platform merging predictive and generative AI. [Business Wire]
InMoment has launched AI Journey Insights, a platform that aims to transform customer experience management by automatically mapping and analyzing journeys across all channels with AI-driven insights. [Business Wire]
Insightsoftware’s business intelligence solution, Logi Symphony, is now available on Google Cloud Marketplace, enhancing data-driven decision-making and analytics capabilities. [GlobeNewswire]
Intuit announced Intuit Assist, a generative AI-powered assistant for QuickBooks that automates financial tasks, offering personalized insights and expert support. [Business Wire]
Leapfrog AI, a new natural language AI assistant by Optilogic, allows Cosmic Frog users to design supply chains effortlessly without technical expertise. [PRWeb]
OpenScholar, an open-source AI system, enables researchers to synthesize vast scientific literature efficiently, offering a cost-effective alternative to proprietary models like GPT-4o. [VentureBeat]
OpenText's latest Cloud Editions 24.4 update expands its AI-driven Aviator platform, enhancing enterprise software with deeper integrations and ROI-focused capabilities. [VentureBeat]
Planview, maker of an end-to-end platform for strategic portfolio management and digital product development, has released new skills in its AI-powered assistant, Planview Copilot. [Business Wire]
Procore Technologies announced the launch of Procore AI solutions, including AI-powered Agents and Insights, to enhance construction project efficiency and decision-making. [Business Wire]
Quik! and Forms Logic have expanded their partnership, using FormXtract to automate and enhance data accuracy in financial services transitions. [PRWeb]
Rezolve Ai and Google have partnered to enhance digital commerce through AI solutions on Google Cloud. [Retail Insight Network]
Robling and AnswerRocket have launched a generative AI retail analytics platform with an AI assistant named Max that provides real-time insights for improved retail decision-making. [Business Wire]
Salesbot.io offers an AI-powered sales platform to automate lead generation and outreach to 655 million contacts, enabling personalized campaigns and faster deal closures. [PRWeb]
Shiprocket has launched Shiprocket Copilot, an AI virtual assistant designed to streamline operations and enhance productivity for Indian sellers in e-commerce. [Analytics India Magazine]
SmartLifeSkills.ai offers a comprehensive platform with AI-powered multilingual life skills training, job resources, and interactive tools to enhance personal and career development. [PRWeb]
Syncfusion announced Bold BI Volume 8, featuring an AI-powered analytics assistant and enhanced features for improved data visualization, performance, collaboration, and platform support. [GlobeNewswire]
Temenos and Nvidia are partnering to deliver generative AI solutions to banks. [FinTech Global]
Tidio's Lyro AI has expanded to email, achieving up to 90% automated resolution, enhancing customer support by integrating across major communication channels. [PRWeb]
And some of Substack’s biggest newsletters rely on AI writing tools. [Wired]
Risks and Responses
AI Unlimited Group announced the formation of an AI Ethics Board and intensified R&D to pioneer ethical AI development. [GlobeNewswire]
Anthropic is working with the US Department of Energy’s nuclear specialists to ensure its models don’t help people make weapons. [Axios]
And Anthropic’s CEO says mandatory safety tests are needed for AI. [Bloomberg]
Character.ai's new safety roadmap claims to enhance protection, particularly for minors, but doubts remain about its enforcement of content moderation and user safety. [Futurism]
CloudSEK's new deepfake detection tool uses a Fakeness Score to help users identify AI-generated content by analyzing inconsistencies in video and audio. [TechRadar]
Factiverse, a Norwegian startup specializing in live fact-checking, aims to combat online disinformation exacerbated by generative AI during the US 2024 election. [TechCrunch]
Google's Gemini AI assistant reportedly issued a disturbing threat to a user. [TechRadar]
The IEEE Standards Association has launched Joint Specification V1.0 to enhance the trustworthiness of AI systems globally, aligning with the EU AI Act. [Business Wire]
The US Patent and Trademark Office banned generative AI for work tasks due to security concerns, but allows controlled testing within the agency to explore potential applications. [Wired]
And the US Department of Homeland Security released a new AI safety framework for critical infrastructure, addressing supply chain risks to ensure security and resilience. [TechRadar]
Elon Musk’s X is suing California over a new law banning AI-generated election misinformation, arguing it infringes on First Amendment rights by censoring political speech. [Engadget]
Regulation
SAS has committed to early compliance with the EU AI Act principles by joining the AI Pact to promote ethical AI governance, literacy, and collaboration. [PR Newswire]
A lobbying group representing the tech industry is urging the Trump administration to review AI regulations, develop international AI standards, and enact a national AI law focusing on risk management. [Semafor]
But Trump may repeal Biden’s AI executive order, leading to state-level regulations, challenges for enterprises, and less government-funded AI research. [VentureBeat]
The EU’s first draft of a Code of Practice for general-purpose AI providers invites feedback until November 28. [TechCrunch]
The UK’s Competition and Markets Authority concluded that Google's partnership with Anthropic doesn’t warrant further investigation, affirming Anthropic’s independence and lack of material influence from Google. [Yahoo Finance]
The US FDA will hold a Digital Health Advisory Committee meeting, discussing AI-enabled medical devices, open for public comments. [US FDA]
Conversational AI
Avaamo has launched AutoQA, a generative AI-powered quality assurance tool for contact centers, enabling comprehensive, accurate evaluation of customer interactions across all channels. [EIN Presswire]
ElevenLabs launched a platform for building customizable conversational AI bots, enhancing integration of knowledge bases and leveraging its text-to-speech expertise. [TechCrunch]
Fisher & Paykel is deploying Salesforce's Agentforce to enhance customer support with autonomous AI agents, allowing human reps to focus on higher-value tasks. [Yahoo Finance]
GR8 People introduced Hugo, an AI chatbot enhancing global recruitment by streamlining candidate interactions and automating job matching and interview scheduling. [EIN Presswire]
Maven AGI's enterprise AI platform unifies fragmented systems, enabling rapid deployment of intelligent agents to enhance customer and employee experiences with accurate, reliable information. [PRWeb]
MediaMonk.ai has launched Phona, a human-like AI voice assistant designed to automate phone call communications and enhance small business client interactions with advanced multilingual capabilities. [EIN Presswire]
Mitel announced its new AI-powered customer experience platform, Mitel CX, designed to enhance customer engagement, employee efficiency, and integrate with existing communications solutions. [Business Wire]
SoundHound AI and Apivia Courtage's partnership successfully utilized Amelia AI Agents to handle over 100,000 customer calls, enhancing service efficiency and employee focus. [Business Wire]
Talkdesk’s Embedded product integrates its AI-powered solutions into popular CRM and helpdesk systems, enhancing agent efficiency and reducing system overwhelm in contact centers. [Business Wire]
A major US travel company replaced its legacy IVR with Verint's AI-powered Intelligent Voice Assistant, increasing self-service call containment to 50% and saving US$10m annually. [Business Wire]
Zoom's AI outperforms Microsoft Teams and Webex, with superior transcription accuracy, meeting summaries, and faster conversational AI responses. [TechRadar]
Be Real
Duolingo's new eyerolling emo chatbot Lily briefly replaced its CEO on an investor call to showcase its AI technology. [Fortune]
ESPN is developing an AI-generated avatar called FACTS to present sports analytics in an engaging way, using Nvidia’s ACE and other tech integrations. [The Verge]
A Swiss chapel uses an ‘AI Jesus’ in confessional, offering 24/7 machine-generated advice. [TechCrunch]
Technology has been used to recreate the voice of the medieval king Richard III, complete with a distinctive Yorkshire accent. [The Guardian]
AI can now create a highly accurate virtual replica of your personality from a two-hour interview. [MIT Technology Review]
UK startup Surf Security launched a beta browser with a tool detecting AI-generated deepfakes with 98% accuracy. [The Next Web]
And Sapia.ai, a provider of AI-powered chat interviewing, has upgraded its AI content detection system to also achieve 98% accuracy. [Business Wire]
Voice News
Israeli startup aiOla has introduced Whisper-NER, an open-source AI model integrating ASR and NER to enhance privacy by masking sensitive speech transcription data. [VentureBeat]
DeepL has launched DeepL Voice for real-time text translations in multiple languages during audio conversations and video conferencing. [TechCrunch]
Google has expanded AI voice capabilities to over a dozen new African languages, enabling voice interactions across Google services for an additional 300 million users. [Slator]
The AINOTE Air 2 e-paper tablet by iFLYTEK offers real-time voice-to-text transcription in 14 languages, multi-language translation, and a natural writing experience. [EIN Presswire]
Mozilla's Common Voice, a volunteer-led effort, aims to diversify AI voice models by collecting multilingual and diverse demographic data from around the world. [MIT Technology Review]
WhatsApp has introduced end-to-end encrypted voice message transcripts, rendered on-device, available globally in select languages, with more to be added in the future. [Engadget]
Translation
KUDO is partnering with Slator to provide live AI translation and captioning services for the SlatorCon Remote event. [Slator]
MotionPoint introduced AdaptiveQE, an automated translation quality estimation technology that reduces costs and maintains accuracy by focusing human review on high-impact content. [PRWeb]
Smartling has announced several significant platform enhancements, including a suite of quality assurance tools, data visualization capabilities, and an innovative AI Translation offering engineered to replace traditional machine translation post editing at half the cost. [MultiLingual]
Unbabel has launched Widn.AI, a standalone AI solution powered by a multilingual LLM that recently outperformed GPT, DeepL, and Google Translate. [MultiLingual]
Veen Bosch & Keuning, the Netherlands’ largest publisher, has announced plans to trial the use of AI to translate a limited selection of books into English, sparking debate. [MultiLingual]
Search
Dnotitia unveiled Seahorse, the world’s first hardware-accelerated vector database, enhancing AI semantic search. [EIN Presswire]
The DOJ plans to urge a judge to force Google to sell Chrome, alongside AI and Android measures, to address its search market monopoly, with a final ruling expected by August 2025. [Verdict]
But Google criticized the DOJ’s proposal, arguing it could harm American tech leadership, increase device costs, and risk consumer privacy. [TechRadar]
And DuckDuckGo urged the European Commission to launch three investigations into Google's compliance with the EU’s Digital Markets Act, citing inadequate enforcement. [Yahoo Finance]
This piece asks whether Google Scholar can survive the AI revolution. [Nature]
NytroSEO has introduced AI-powered user search intent optimization to transform Google search optimization. [MarTech Series]
RavenPack’s Bigdata.com, powered by Vespa.ai, revolutionizes financial research by enabling real-time interaction with billion-scale vector search, enhancing decision-making with speed and accuracy. [Business Wire]
Health Tech
Academy Medtech Ventures has launched Move PT, an AI-driven digital health platform for personalized rehabilitation, improving outcomes through real-time data and analytics. [Business Wire]
A survey by Elation Health reveals 66% of primary care physicians view AI as reducing administrative burdens, improving outcomes, and enhancing practice efficiency. [Business Wire]
G2 Speech has launched SpeechReport 4.0, a major upgrade utilizing AI to enhance medical documentation workflow, offering 99% accuracy and extensive customization features. [Health Tech World]
IntelePeer and 42 North Dental have partnered to deploy AI-powered solutions to enhance patient communications, improve efficiency, and streamline dental operations. [Business Wire]
Meta aims to integrate its Llama AI into the UK’s public healthcare system, focusing on improving services. [Engadget]
ModMed’s AI-powered Scribe aims to transform healthcare by reducing administrative burdens and improving patient experiences. [Business Wire]
PredictionHealth has partnered with FYZICAL Therapy & Balance Centers to provide AI scribing and analytics, enhancing operations and patient care across 600 clinics nationwide. [PRWeb]
Suki AI announced partnerships with MEDENT and Azalea Health to integrate its AI engine into their EHR systems. [Business Wire]
Central Virginia Health Services uses Sunoh.ai to enhance clinical documentation. [Business Wire]
Thoughtful AI launched specialized AI Agents for revenue cycle management, enhancing healthcare efficiency, accuracy, scalability, and reducing costs while improving patient care. [Business Wire]
Thrive AI Health, backed by Thrive Global and the OpenAI Startup Fund, aims to develop an AI health coach for personalized wellness advice. [TechCrunch]
But TechRadar finds that Thrive AI Health appears underdeveloped, with limited unique features. [TechRadar]
Veradigm has launched Veradigm Ambient Scribe, an AI-enabled platform that eases healthcare providers’ administrative tasks by transcribing patient interactions into structured notes. [Business Wire]
Legal Tech
DigitalOwl and ExamOne are collaborating to enhance medical record retrieval and analysis for insurance and legal professionals through advanced AI solutions. [Business Wire]
Eon Media has launched GovSessions Insights, an AI platform using ML and NLP to transform professional interactions with legislative sessions. [PRWeb]
Icertis and Microsoft have partnered to integrate AI-powered contract intelligence into Dynamics 365 Supply Chain Management, enhancing supplier management, cost control, and risk mitigation. [Business Wire]
The International Legal Technology Association’s Litigation Special Interest Group released a comprehensive guide for using Active Learning in eDiscovery to standardize legal disclosure. [Artificial Lawyer]
Ironclad launched Jurist, an AI-powered legal assistant that automates and streamlines legal work by utilizing multi-agent technology within a centralized, privacy-focused platform. [AiThority]
Knowable announced Ask Knowable, AI-powered tools transforming contract management by utilizing generative AI and accurate data to enhance understanding and interaction with executed agreements. [LawSites]
LexisNexis launched the Lexis+ AI Mobile App, allowing legal professionals to access AI-assisted legal research and drafting capabilities on mobile devices. [LawSites]
Meanwhile, Professor Benjamin Perrin criticized Lexis+ AI for inaccuracies in legal research tasks, with LexisNexis defending its tool and emphasizing ongoing improvements and the need for better user communication. [Artificial Lawyer]
Pramata launched its AI Design Studio, an AI-powered contract management tool that enhances contract automation, insights, and strategic efficiencies for legal and business teams. [Business Wire]
ProSearch has partnered with eDiscovery AI to enhance document review efficiency and precision by integrating advanced AI technology with ProSearch’s data management expertise. [KMWorld]
Thomson Reuters is introducing its AI-driven CoCounsel for Drafting in the UK, using Practical Law data for enhanced contract drafting. [Artificial Lawyer]
Trellis has launched Trellis AI, a suite of generative AI tools leveraging extensive litigation data to automate and enhance productivity in trial court tasks like motion drafting and case analysis. [LawSites]
Funding
Amazon will invest US$110m in university-led generative AI research, providing resources like Trainium chips to support academic innovation and future AI expertise. [IEEE Spectrum]
Arphie raised a US$2.9m seed round to enhance collaborative intelligence in B2B software, pioneering AI-driven RFP and questionnaire automation. [Arphie]
Swiss startup Calvin Risk raised US$4m to enhance AI risk assessment tools and expand operations, ensuring compliance with EU AI regulations for safe AI deployment. [Tech.eu]
Citizen Health secured US$14.5m in Seed funding to advance patient health data solutions and drug development acceleration. [FinSMEs]
CommBox, an Israeli AI-powered customer experience platform provider, received US$15m to expand in North America and Europe and enhance AI developments. [FinSMEs]
Cresta, an AI platform for contact centers, received funding from Accenture Ventures to expand operations and will integrate its software into Accenture’s AI Refinery. [FinSMEs]
Thrive Capital is in talks to lead a US$1B tender offer for analytics firm Databricks, potentially valuing it at US$55B. [Verdict]
Distyl raised US$20m in Series A funding to enhance its AI solutions for Fortune 100 companies. [PR Newswire]
Google is launching a US$22m initiative through Google.org to fund scientists and researchers using AI for scientific breakthroughs, positioning itself as a key partner for innovative projects. [TechCrunch]
The Abu Dhabi sovereign wealth fund has acquired a stake in the US software company Qlik. [The Financial Times]
Revisto, an AI-powered platform improving medical, legal, and regulatory review for pharmaceuticals, raised US$4m. [FinSMEs]
Rox, which is developing AI-powered agents for salespeople, raised US$50m in seed funding. [Bloomberg]
runQL, a Waterloo-based tech startup, raised US$1.6m in Pre-Seed funding to enhance its smart query platform for data professionals. [FinSMEs]
Seasalt.AI, a Seattle startup whose software can answer questions automatically via an AI agent with either voice or text, announced a US$4.2m seed round. [GeekWire]
Spines raised US$16m to revolutionize publishing with AI, expanding operations into audiobooks and global markets. [Refresh Miami]
SuperAnnotate raised US$36m in Series B funding to enhance its enterprise software for AI dataset management. [FinSMEs]
Bulgarian startup Team-GPT raised US$4.5m to help enterprises securely leverage AI in collaborative workspaces. [AiThority]
Theo Ai, a predictive AI platform for litigation leveraging generative AI, raised US$2.2m pre-seed funding aiming to accurately predict case outcomes. [Artificial Lawyer]
Wordware secured a US$30m seed round to simplify AI development using natural language, aiming to empower domain experts over traditional engineers. [VentureBeat]
Acquisitions
Anysphere acquired AI coding assistant Supermaven to enhance Cursor’s capabilities, integrating Supermaven’s technology and team to create a more advanced, user-friendly AI code editor. [TechCrunch]
ChapsVision has acquired Sinequa for AI-powered enterprise search, backed by a US$90m investment. [Business Wire]
Front has acquired Idiomatic to enhance its AI-powered customer service platform, leveraging advanced VoC intelligence to improve customer insights and drive smarter, data-driven decisions. [Business Wire]
There’s More
In 2017, OpenAI considered acquiring AI chipmaker Cerebras Systems via Tesla, but ultimately abandoned those plans, opting to work with Broadcom and TSMC on in-house chips. [TechCrunch]
The Rabbit R1 can now use AI to create customized interfaces. [The Verge]
Quantization can make AI models more efficient but has limitations, especially for large models, and extremely low precision may harm model quality. [TechCrunch]
James and Rose, AI news broadcasters at The Garden Island, were terminated due to negative public response and resource concerns after a two-month run. [Wired]
KPMG plans to invest US$100m in AI services through Google Cloud, focusing on development and employee training, as AI transforms professional services. [Yahoo Finance]
The Back Cover
Here’s Jeff Beck with a cover of Charlie Mingus’ Goodbye Pork Pie Hat.
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!