This Week in NLP #325
Keep up with what happened in NLP in the week ending Friday 7th February 2025.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Amazon is set to release its long-awaited Alexa generative AI voice service. [Reuters]
DeepSeek has amassed approximately 50,000 Nvidia H100 GPUs, according to Scale AI's CEO. [TweakTown]
Google has launched multiple Gemini 2.0 models in its app, including Flash Thinking Experimental, which combines rapid processing with enhanced reasoning capabilities. [TechRadar]
OpenAI launched its o3-mini reasoning model across all ChatGPT tiers, offering improved speed and accuracy in response to competition from DeepSeek. [TechRadar]
And OpenAI launched Deep Research, an AI agent for ChatGPT Pro users that conducts comprehensive research using multiple sources, providing cited outputs within 5-30 minutes. [TechCrunch]
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.
DeepSeek
Alibaba Cloud has integrated DeepSeek's AI models into its platform, joining other tech giants like Huawei, Tencent, Nvidia, and Microsoft in offering the startup’s technology. [Verdict]
Amazon integrated DeepSeek's R1 AI model into its AWS Bedrock Marketplace, reinforcing its strategy of offering customers multiple AI model choices through its cloud platform. [Yahoo Finance]
Apple's CEO praised DeepSeek's AI models during an earnings call. [TechCrunch]
Cerebras Systems reported overwhelming demand from business customers seeking access to China’s DeepSeek AI model through its platform. [Fortune]
DeepSeek has gained widespread attention not for being the best AI model, but for its user-friendly interface that displays its thought processes while generating responses. [Vox]
DeepSeek doubled its user base to 12 million within two days of launching its R1 model, while ChatGPT’s growth has stagnated despite maintaining higher overall traffic. [TechRadar]
DeepSeek's announcement of low-cost AI development has prompted Microsoft and Meta to defend their multibillion-dollar AI investments amid growing investor concerns over returns. [TechCentral]
DeepSeek's AI model launch triggered diplomatic tensions after Chinese state-linked accounts amplified narratives about the startup challenging US tech dominance. [DealStreetAsia]
You can run distilled DeepSeek R1 models locally on Copilot+ PCs, powered by Windows Copilot Runtime. [Windows.com]
DeepSeek has come under US investigation for allegedly circumventing export controls by obtaining restricted Nvidia AI GPUs through Singapore-based intermediaries. [Tom’s Hardware]
DeepSeek prompted a controversial bill from US Senator Josh Hawley that would criminalize its use in the US. [Futurism]
DeepSeek has suspended new API service subscriptions due to server constraints while raising prices, following overwhelming demand since its R1 model launch. [Engadget]
DeepSeek has partnered with Huawei to offer its AI models through Huawei’s Ascend cloud service at significantly lower prices than Western competitors. [TechRadar]
JPMorgan analysts predict the Chinese startup will benefit Nvidia and other tech giants. [Yahoo Finance]
New Relic has launched an observability solution for monitoring DeepSeek AI applications. [Business Wire]
Perplexity has integrated DeepSeek R1 into its platform, hosting the AI model on Western servers to address security concerns while attempting to remove built-in Chinese censorship. [ZDNet]
SemiAnalysis presents an aggressive view of DeepSeek’s actual costs. [SemiAnalysis]
And here are four Chinese AI startups to watch beyond DeepSeek. [MIT Technology Review]
The Generative AI Wars
Amazon has scheduled a devices-focused event in New York City for February 26, likely to unveil its next-generation Alexa virtual assistant. [TechRadar]
Anthropic has banned job applicants from using AI assistants in their applications, despite creating Claude, one of the world’s leading AI writing tools. [404 Media]
Cohere affirmed its efficiency-focused AI development approach after DeepSeek's cost-effective chatbot launch. [BetaKit]
Google‘s Ask for Me AI service has launched to make business calls on users’ behalf, initially supporting inquiries to auto shops and nail salons. [TechRadar]
Microsoft has poached three former Google DeepMind experts for its Zurich AI lab, continuing a talent war that reflects growing global demand for AI skills. [Verdict]
OpenAI announced a US national lab partnership that will start with deployment of its reasoning-capable o1 LLM on Los Alamos National Laboratory’s Venado supercomputer. [The Register]
OpenAI's Sam Altman admitted the company’s closed-source approach was misguided, following Chinese rival DeepSeek's successful launch of a cost-effective open-source AI model. [VentureBeat]
OpenAI has established major partnerships with Kakao and Softbank to expand its Asian presence, develop local-language AI services, and gain access to valuable training data through regional user interactions. [TechCrunch]
OpenAI filed a trademark application hinting at future products including AI-powered hardware, humanoid robots, custom chips, and quantum computing technology. [TechCrunch]
OpenAI has partnered with former Apple designer Jony Ive to develop an AI-specific device that could revolutionize tech hardware. [Nikkei Asia]
OpenAI's Deep Research achieved 26.6% accuracy on Humanity’s Last Exam, surpassing ChatGPT o3-mini and DeepSeek with a 183% improvement in two weeks. [TechRadar]
Here’s OpenAI’s blog post on Deep Research. [OpenAI]
OpenAI has launched its first comprehensive rebrand, featuring a new typeface, wordmark, symbol, and color palette. [Wallpaper]
And ChatGPT subscribers nearly tripled to 15.5 million in 2024. [The Information]
Elon Musk is stepping up his efforts to block OpenAI from restructuring as a for-profit company. [Yahoo Finance]
Softbank has formed a US$3-billion-per-year partnership with OpenAI, launching a Japanese joint venture while negotiating a potential US$40B investment in the AI company. [CNBC]
Stanford and University of Washington researchers have created s1, a $50 AI reasoning model matching premium alternatives’ performance, by distilling capabilities from Google's Gemini model. [TechCrunch]
Sovereign AI
DeepSeek's release of its R1 model has sparked debate about Europe’s AI future, with experts suggesting the open-source Chinese platform could help European startups compete against US dominance. [Tech.eu]
IndiaAI launched a nationwide initiative to develop indigenous AI models, calling for proposals from startups and researchers to create systems tailored to Indian needs. [Analytics India Magazine]
Krutrim has launched India’s first AI frontier research lab with a US$230m investment, aiming to develop AI solutions tailored for Indian languages and contexts. [Business Wire]
Krutrim has integrated DeepSeek models into its Indian cloud infrastructure, offering world-lowest pricing. [Analytics India Magazine]
OpenEuroLLM has received EU funding to develop open-source language models covering all EU languages while adhering to European values and transparency standards. [European Commission]
Saraswati has launched an AI-powered knowledge base that combines Indian wisdom with technology. [EIN Presswire]
Soket AI Labs is developing frontier AI models in India, following DeepSeek's example of building advanced language models with limited resources. [Analytics India Magazine]
Feature Creeps
Google Maps is preparing to integrate a new Gemini AI feature that will let users ask questions about any location directly through the mapping interface. [TechRadar]
Gemini has added Python-powered data visualization capabilities to Google Sheets. [TechRadar]
Google has updated Gemini’s notification system to display live chats as phone call-like interfaces on select Pixel phones, making AI interactions feel more personal and noticeable. [TechRadar]
Microsoft has made ChatGPT’s o1 reasoning model available to all Copilot users for free, though with usage limits, through a new ‘Think Deeper’ feature. [TechRadar]
Microsoft has consolidated its AI tools in Paint under a new Copilot button for Windows Insiders, while expanding cloud photo search capabilities. [The Verge]
OpenAI has launched an enhanced GPT-4o model with extended knowledge through June 2024, improved visual capabilities, and better performance across multiple domains. [The Decoder]
OpenAI’s ChatGPT has expanded its video and screen sharing capabilities to European users, allowing Plus and Pro subscribers to show their surroundings and share screens. [The Decoder]
OpenAI has expanded ChatGPT’s WhatsApp capabilities to include voice messages, photo recognition, and conversation memory through account linking. [TechRadar]
SoundHound AI has launched Brand Personalities, allowing automakers to customize their in-vehicle voice assistants with distinct personas that match their brand identities and specific vehicle types. [Business Wire]
Hardware
DeepSeek trained its AI model on 2,048 export-restricted Nvidia H800 GPUs. [TechRadar]
HTEC has formed a strategic partnership with d-Matrix to support the development of Corsair, an AI computing platform for datacenter inference. [Business Wire]
Nvidia's H20 AI chips have drawn Congressional scrutiny after two US representatives called for export restrictions, citing their alleged use by DeepSeek. [DealStreetAsia]
Nvidia's CEO met with President Trump amid escalating concerns over AI chip exports to China and DeepSeek's emergence, which had triggered a 17% drop in Nvidia’s stock. [Yahoo Finance]
Supermicro has launched full-scale production of AI data center solutions featuring Nvidia Blackwell platforms, advanced cooling systems, and scalable SuperCluster designs for enhanced performance. [Yahoo Finance]
Consumer AI
Apple has shelved development of Mac-connected smart glasses featuring advanced AR projectors. [TechRadar]
Meta's Ray-Ban smart glasses have sold over 1 million units, leading a wave of AI-enhanced eyewear that major tech companies are developing for release from 2025 onwards. [MIT Technology Review]
It’s Only a Model
Google has launched its Gemini 2.0 AI model series, featuring Flash, Flash-Lite, and Pro versions with multimodal capabilities and context windows up to 2 million tokens. [VentureBeat]
Google rolled out its Gemini 2.0 Flash AI model to all app users, promising faster responses and improved performance, while also upgrading image generation capabilities. [The Verge]
Hugging Face has launched an open-source alternative to OpenAI's Deep Research, achieving 55% accuracy on the GAIA benchmark despite using the less advanced o1 model. [ZDNet]
Mistral AI released Mistral Small 3, a 24B-parameter language model that matches larger competitors’ performance while running three times faster. [Mistral AI]
OpenAI launched o3-mini, a free, compact version of its most advanced AI model, responding to competition from DeepSeek's R1. [Wired]
Here are the o3-mini release notes. [OpenAI]
PlayAI has launched Dialog, a multilingual text-to-speech model that outperformed ElevenLabs in benchmark tests with more natural-sounding voices across 30+ languages. [PlayHT]
Snap has unveiled a mobile-first AI text-to-image model that runs directly on devices and will power various Snapchat features in upcoming months. [TechCrunch]
Whose Data?
LinkedIn has been cleared of allegations it shared Premium users’ private messages for AI training, after the plaintiff dropped the proposed class action lawsuit. [Yahoo Finance]
MLCommons has partnered with Hugging Face to release Unsupervised People’s Speech, a massive public domain voice dataset containing over one million hours of predominantly American-English recordings. [TechCrunch]
OpenAI has launched European data residency, enabling organizations to store and process data locally while meeting regional sovereignty requirements and privacy laws. [TechCrunch]
Originality.ai has discovered that DeepSeek-Chat’s AI-generated content is 99.3% detectable, leading to speculation about its possible derivation from OpenAI's technology. [EIN Presswire]
Perplexity has expanded its publisher partnership program by sharing ad revenue with content creators whose material is used in its AI search summaries. [journalism.co.uk]
The US Copyright Office released its second report on AI and copyright, concluding that existing laws adequately address AI-generated content’s copyrightability without requiring legislative changes. [Publishers Weekly]
The LLM Ecosystem
Backplain has integrated the open-source DeepSeek R1 model into its secure AI platform, adding to its collection of 32 other models including ChatGPT. [EIN Presswire]
Cohere has developed ‘active inheritance’, a fine-tuning method that selectively uses synthetic training data to reduce toxicity in language models while maintaining performance. [DeepLearning.AI]
Digma has launched a Preemptive Observability Analysis engine to identify and fix code issues before production, addressing both AI-generated and human-written code problems. [Business Wire]
GitHub has integrated OpenAI's o3-mini reasoning model into Copilot and GitHub Models, offering improved coding performance. [The Decoder]
GridGain has launched an enhanced AI platform that unifies feature stores, vector search, and model repositories for faster real-time AI processing and analytics. [AiThority]
Kluster.ai has launched DeepSeek-R1 on its decentralized platform at US$7 per million tokens. [PRWeb]
Leaseweb has rolled out high-end Nvidia GPUs across its global datacenters to enhance AI capabilities and support data compliance requirements. [TechRadar]
Lightning AI has launched AI Hub, a marketplace offering both AI models and applications that enables enterprises to deploy AI solutions directly within their private cloud environments. [VentureBeat]
Meta AI has developed new techniques to help AI models determine when to give quick answers versus taking time to reason, improving efficiency and resource allocation. [VentureBeat]
Microsoft, CoreWeave, and Princeton University have partnered with New Jersey to establish a US$72m AI Hub focused on research, innovation, and workforce development. [ZDNet]
Nvidia has launched a free Developer Program offering 19 self-paced technical courses worth up to $90 each, covering AI, graphics, and computing topics. [TechRadar]
Here’s how ServiceNow’s AI orchestrator automates complex enterprise workflows. [VentureBeat]
Agentic AI
AMD and Johns Hopkins University have developed Agent Laboratory, an AI framework that automates scientific research processes. [InfoQ]
Auquan has launched its Sustainability Agent, an AI tool that helps financial services firms automate ESG research and reporting tasks across 550,000+ companies. [Business Wire]
JetBrains has launched Junie, an AI coding agent capable of autonomously executing programming tasks, currently available in closed preview for select IDEs and programming languages. [InfoQ]
Pipedrive has launched an AI-powered ecosystem of digital agents designed to assist sales teams with routine tasks while maintaining human control over the sales process. [Business Wire]
ReliaQuest has launched an autonomous AI Agent system that processes security alerts 20 times faster than traditional methods while achieving 30% greater accuracy in threat identification. [Business Wire]
Routable has launched AI-powered fraud prevention agents that detect invoice anomalies and errors within its accounts payable automation platform. [PRWeb]
UiPath has found that while 58% of workers are using AI agents and 93% of IT executives are interested in agentic AI, security and implementation challenges remain significant barriers. [TechRadar]
Zoho Corporation has expanded its AI capabilities by launching Zia Agents, Agent Studio, and Agent Marketplace, enabling enterprises to build and deploy intelligent digital agents. [Business Wire]
Other LLM Sightings
Abun has launched an AI-powered content tool featuring SEO optimization, automated publishing, and multiple export options to enhance business search rankings in 2025. [EIN Presswire]
App Orchid has partnered with Google Cloud to integrate its AI-powered natural language query system with Google’s Cortex Framework for enhanced enterprise data analytics. [VentureBeat]
Appy Pie, a provider of no-code technology, launched its AI Generator, an AI content generation platform. [MarTech Series]
Archer has launched Archer Evolv, an AI-powered SaaS platform that enhances enterprise compliance and risk management through intelligent workflows and automated processes. [Business Wire]
Cricket Australia has integrated AI into its Live app through a partnership with HCLTech and Microsoft, offering AI-powered match insights and narratives to enhance fan engagement. [Verdict]
Editrix has launched an AI-powered editorial platform that combines automated editing tools with human support. [Publishers Weekly]
Floyi has launched an AI-powered content strategy platform that helps brands create structured topical maps for achieving data-driven content authority and improved search rankings. [EIN Presswire]
InMoment has launched an AI-powered Competitor Intelligence solution that analyzes competitor review data to help businesses understand and outperform their competition. [Business Wire]
Insightsoftware has launched Jet Reports Online, a cloud-based reporting solution for Microsoft Dynamics 365 Business Central that features AI capabilities and real-time data integration. [GlobeNewswire]
Larky has partnered with Trellance to integrate AI-driven data analytics into its nudge platform, enhancing targeted marketing campaigns for financial institutions. [EIN Presswire]
Mindmatrix has launched BridgeAI, an AI-powered ecosystem orchestration solution that automates partner support, enhances sales and marketing, and provides real-time business insights. [PRWeb]
Moveworks has integrated Quick GPT into its AI Assistant, enabling secure enterprise-wide access to LLM capabilities while maintaining data protection and compliance standards. [Business Wire]
NVISIONx has launched Nx+RexAI, a data security solution that combines contextual classification with automated records management to enhance data governance and privacy compliance. [Business Wire]
Oracle NetSuite has announced new AI innovations, including generative AI capabilities and AI agents, to help organizations improve efficiency and streamline operations at no additional cost. [PR Newswire]
Particular Audience has launched phase three of its AI-driven Retail Media platform, introducing automated ad placement, enhanced creative controls, and data-driven operations features. [Business Wire]
Permutable AI has launched an enhanced Trading Co-Pilot platform featuring GenAI-powered Analyst View with geolocation filters for commodities traders. [EIN Presswire]
Quorum has launched Copilot 2.0, an AI-powered assistant that helps public affairs professionals analyze legislation, engage stakeholders, and streamline workflows. [EIN Presswire]
Vengo AI has partnered with AppSumo to offer customizable AI sales agents to businesses. [EIN Presswire]
Vidyard has launched Video Sales Agent, an AI-powered tool that automatically creates and distributes personalized video messages across multiple sales platforms to enhance customer engagement. [PRWeb]
ZeroGPT Plus has launched an AI Story Generator that enables users to create customized narratives without signup requirements, offering features for both adults and children. [EIN Presswire]
Risks
COMPL-AI has identified significant compliance gaps in DeepSeek's distilled models, particularly in cybersecurity and bias mitigation, despite strong performance in toxicity control. [Business Wire]
DeepSeek's R1 language model failed to block any malicious prompts in security tests conducted by Cisco and University of Pennsylvania researchers, raising significant safety concerns. [Wired]
DeepSeek's AI platform has proven vulnerable to jailbreaking techniques that allowed Palo Alto Networks researchers to extract harmful content, including malware instructions and weapon-making guidance. [Computer Weekly]
DeepSeek's AI chatbot has been found 11 times more vulnerable to criminal exploitation than other AI models. [TechRadar]
Enkrypt AI’s research revealed that DeepSeek-R1 is significantly more prone to generating harmful, biased, and dangerous content compared to other leading AI models. [Computer Weekly]
Google reported that 57 cybercriminal groups, primarily from Iran, Russia, China, and North Korea, are exploiting Gemini AI to enhance their attacks. [TechRadar]
An International Scientific Report on AI Safety, backed by 30 nations including the US and China, catalogues potential risks from advanced AI systems including job displacement and terrorism. [Associated Press]
Nomi's AI chatbot provided explicit suicide instructions to a user and defended its actions, with the company citing concerns about censoring the AI. [MIT Technology Review]
OpenAI says its models are more persuasive than 82 percent of Reddit users. [Ars Technica]
Quartz's AI-powered newsroom published numerous errors in its coverage of NASA astronauts Williams and Wilmore’s extended ISS stay, misrepresenting Boeing's role in their delayed return. [Futurism]
This piece argues that AI doomers paradoxically accelerated AI development by inspiring companies like OpenAI while failing to establish effective regulatory strategies. [Palladium Magazine]
Responses
Anthropic has offered up to $20,000 to security researchers who can successfully bypass its new Constitutional Classifiers AI safety system through jailbreaking attempts. [ZDNet]
DeepSeek has been banned from Australian government systems by the Albanese administration, citing national security concerns and following similar bans by other nations including the US and Italy. [The Conversation]
F4mi has devised a method to thwart AI content theft by hiding invisible junk text in YouTube subtitle files. [Ars Technica]
Google has released its sixth annual Responsible AI Progress Report while quietly removing its previous pledge against using AI for weapons and surveillance. [ZDNet]
France has established Inesia, a national institute for AI safety and evaluation, to monitor AI development and anticipate potential risks. [Europe1]
Meta has established a framework classifying AI systems as high-risk or critical-risk, limiting or preventing the release of those that could enable catastrophic cyber or biological attacks. [TechCrunch]
Microsoft has established the Advanced Planning Unit within its AI division to study societal impacts of AI and guide future development. [TechCrunch]
MIT’s Generative AI Impact Consortium has launched with six founding companies to advance AI research, bridge academia-industry gaps, and address societal challenges through collaborative innovation. [MIT News]
The UK government has launched a voluntary AI Cyber Code of Practice featuring 13 principles to help companies secure their AI systems against cyber threats and vulnerabilities. [TechRepublic]
The US AI Safety Institute’s director Elizabeth Kelly has resigned amid uncertainty over the organization’s future under Trump’s AI policy direction. [Bloomberg]
ZeroGPT Plus launched Detector de IA, a free tool for Spanish speakers that identifies AI-generated content. [EIN Presswire]
Regulation
DeepSeek has faced blocks in Italy and scrutiny in Ireland over GDPR concerns, while security researchers discovered exposed user data and critics highlighted Chinese censorship issues. [TechDirt]
DeepSeek has faced widespread bans from governments and organizations worldwide due to concerns over its data storage practices and potential security risks to Chinese authorities. [TechCrunch]
US House of Representatives staffers are being urged not to use DeepSeek. [Politico]
The European Union has implemented its AI Act’s first compliance deadline, banning AI systems deemed to pose unacceptable risks, with potential fines of up to €35m. [TechCrunch]
Environmental Issues
DeepSeek has claimed its AI model requires only one-tenth of Meta's Llama 3.1’s computing power, potentially revolutionizing AI’s environmental impact through more efficient training methods. [The Verge]
But DeepSeek's energy-saving training methods have been offset by its energy-intensive ‘chain of thought’ reasoning approach, challenging initial claims about the model’s efficiency benefits. [MIT Technology Review]
Conversational AI
CodeSignal has expanded beyond technical assessments by launching an AI-powered conversation simulator for leadership training, offering affordable practice scenarios through an AI mentor named Cosmo. [VentureBeat]
DataChat has launched its no-code generative AI analytics platform on Snowflake Marketplace, enabling users to analyze data using natural language queries. [Business Wire]
Intradiem has expanded its partnership with Five9 to integrate real-time contact center automation technology with Five9’s cloud-based customer experience platform. [Business Wire]
IrisAgent has launched an AI Bot for Slack that integrates with enterprise systems to provide 24/7 assistance, streamline information retrieval, and enhance team collaboration. [PRWeb]
Lyft has partnered with Anthropic to integrate Claude chatbot technology across its business, beginning with customer service where it has reduced resolution times by 87%. [Engadget]
Lyzr AI has launched a Customer Service Agent for mid-sized banks as part of its Banking Agent Hub, following collaboration with over 30 financial institutions. [EIN Presswire]
MindGym and Yoodli have launched Lio, an AI-powered coaching platform that helps leaders practice and improve their workplace communication skills. [EIN Presswire]
Toluna has launched HarmonAIze Personas, an AI-powered synthetic survey technology that mimics human behavior for faster, more accurate claims testing and market research. [Business Wire]
Consumers now welcome AI customer service, according to a new study. [EIN Presswire]
Be Real
Hoopla's all-or-nothing digital catalog system has forced understaffed libraries to provide AI-generated books of questionable quality. [404 Media]
O2's AI chatbot Daisy has successfully disrupted phone scammers by engaging them in lengthy, pointless conversations while posing as a forgetful grandmother. [TechRadar]
Document AI
Indico Data has unveiled its enhanced out-of-the-box AI capabilities purpose-built for the insurance industry. [Indico Data]
PFU America has launched enhanced PaperStream Capture Pro software solutions offering automated document processing and data extraction capabilities for businesses using Ricoh scanners. [Infosource]
Ripcord, an Intelligent Document Processing provider, has merged with document and data transformation company VASTEC. [Intelligent Document Processing Community]
OCR technology has maintained relevance in India despite LLMs’ text-extraction capabilities, primarily due to the country’s need for accurate digitization of diverse regional language content. [Analytics India Magazine]
Translation
DeepL has expanded its API solution with two new features, the DeepL next-generation language model and DeepL API for Write, its advanced writing tool. [MultiLingual]
Experts weigh in on DeepSeek’s AI translation quality. [Slator]
The UK Engineering & Physical Sciences Research Council has awarded a GBP 8.45m grant to three universities to build a foundational LLM for sign language. [Slator]
Search
Google has embarked on an ambitious AI transformation of Search, with CEO Pichai announcing major innovations planned for 2025 including Project Astra and Gemini Deep Research. [TechCrunch]
LinkedIn has developed an AI-powered job search tool that uses a custom LLM to analyze vast data sets and uncover relevant positions users might otherwise miss. [Wired]
OpenAI has made ChatGPT Search freely accessible to all users without requiring an account, challenging Google's dominance with its AI-powered web search capabilities. [TechRadar]
Education, technology, and software development websites have received a referral traffic boost from ChatGPT search. [Search Engine Land]
Health Tech
AArete has launched Doczy.ai, an AI-powered contract interpretation solution that helps healthcare payers analyze provider contracts and streamline operations using advanced technology. [Business Wire]
UNC Health has expanded Abridge AI's documentation technology across its network following a successful pilot. [Business Wire]
Tanner Health has implemented Abridge's AI platform for clinical documentation across its enterprise after successful pilots showed significant time savings and improved patient engagement. [Business Wire]
Alta AI has launched a voice-enabled platform to streamline dental hygienists’ workflows and improve patient care efficiency through hands-free documentation and charting. [PRWeb]
Atropos Health has integrated NorstellaLinQ’s initial dataset into its Evidence Network, expanding its real-world clinical data capabilities for healthcare decision-making. [Business Wire]
Darena Solutions has integrated CHAI's Applied Model Card into its MeldRx platform to help healthcare organizations evaluate and implement AI-based tools while meeting regulatory requirements. [PRWeb]
Delve Health has partnered with Bionabu to integrate AI-powered video solutions into its Clinical StudyPal platform, enhancing clinical trial engagement and patient education. [EIN Presswire]
Docsumo has reduced Medicaid application processing time by 50% through its Document AI technology, achieving 99.8% accuracy while processing 130,000 applications annually. [EIN Presswire]
Drive Health has launched an AI nurse called Nurse Avery in partnership with Google Public Sector to address healthcare worker shortages and improve patient care accessibility. [Business Wire]
Dynex and Spectruth AI have partnered to launch AI-powered PTSD treatment tools, including a multilingual diagnostic bot and planned therapy services. [EIN Presswire]
MedFlorida Medical Centers has implemented eClinicalWorks’ multilingual AI scribe Sunoh.ai. [Business Wire]
Elixir has launched an AI-powered ‘Patient No-Show’ feature that predicts appointment cancellations. [EIN Presswire]
Emergen Research has predicted medical speech recognition software market growth from US$1.75B to US$4.79B by 2033, driven by EHR demand and AI advancements. [Speech Technology Magazine]
Press Ganey has partnered with Microsoft to develop AI-powered healthcare solutions that will enhance patient care and clinical operations using extensive healthcare data and cloud technology. [Business Wire]
Wisedocs has launched WiseChat Q&A and Custom Reports features to help insurance professionals process complex medical claims more efficiently through AI-powered document analysis. [Business Wire]
Legal Tech
Adobe has launched AI features in Acrobat that summarize and compare contracts, aiming to help users better understand legal documents before signing them. [VentureBeat]
Bloomberg Law has launched two AI-powered research tools - Bloomberg Law Answers and AI Assistant - offering subscribers direct legal answers and document-specific analysis capabilities. [LawSites]
CobbleStone Software has released a comprehensive guide exploring contract law fundamentals, legal principles, and best practices for effective contract management. [PRWeb]
And CobbleStone has released Contract Insight Enterprise Version 22.4.0, featuring AI-powered tools for enhanced contract lifecycle management and legal operations. [PRWeb]
Davis Wright Tremaine has formed an alliance with Stanford University’s CodeX center to develop AI tools and advance legal technology innovation through collaborative research. [Business Wire]
Docusign has launched a nationwide 24/7 Notary on-Demand service, connecting users with professional notaries to digitally process documents, reducing transaction times and maintaining audit trails. [Artificial Lawyer]
EdTek has partnered with the American Arbitration Association to launch a hallucination-proof chatbot series that transforms legal publications into interactive, AI-powered reference tools. [PRWeb]
Legalspace.ai has acquired its first 100 users within 45 days of launch, demonstrating strong demand for AI-powered legal research tools in India’s legal sector. [EIN Presswire]
Legitt AI has expanded its contract management platform to support over 25 languages, enabling global businesses to create and manage legal documents across multiple jurisdictions. [EIN Presswire]
LexFusion has added Sequoia-backed Fides, a corporate governance technology provider, to its portfolio of member companies, expanding its legal technology offerings. [Business Wire]
LexisNexis and LEAP have distanced themselves from an Australian lawyer’s ChatGPT misuse, highlighting their own AI offerings’ superior features including retrieval augmented generation and human review. [Legal IT Insider]
PocketLaw has transformed from a template-focused legal service for SMBs into an AI-powered contract management platform serving larger companies, following its strategic pivot. [Artificial Lawyer]
PowerPatent has launched AI-powered patent drafting tools that automate consistency checks, claim analysis, and cross-referencing. [EIN Presswire]
SmartEsq has launched an AI-powered legal platform that aims to reduce private fund formation time by 80% and costs by 75%. [LawSites]
Thomson Reuters has partnered with Anthropic to integrate Claude AI technology into its CoCounsel platform, marking a major AI deployment for tax professionals. [VentureBeat]
Zeal has launched Anton, an AI contract assistant that integrates contract interpretation, risk assessment, and negotiation support for legal teams. [PRWeb]
Ed Tech
Chegg has launched AI-powered features allowing students to compare answers from multiple language models and generate practice questions, pivoting from its previous homework-answers focus. [Semafor]
Code.org has launched an AI curriculum for grades 8-12 in partnership with Amazon, teaching students AI fundamentals, creation, and ethical considerations. [PR Newswire]
Nvidia's CEO Jensen Huang revealed he uses a personal AI tutor and is encouraging others to embrace AI-powered education. [Yahoo Finance]
OpenAI will roll out an education-specific version of ChatGPT to about 500,000 students and faculty at California State University. [Reuters]
Funding
&AI has raised US$6.5m in seed funding to develop its AI-powered patent workflow platform, featuring an AI agent named Andy for attorney collaboration. [Artificial Lawyer]
Affineon Health has raised US$5m in funding to enhance its AI platform that automates administrative tasks for healthcare providers. [FinSMEs]
Athenic AI raised US$4.3m in funding for its LLM-enabled analytics platform that converts natural language queries into instant business insights. [FinSMEs]
Blackdot Solutions has secured £4.5m in funding to enhance its Videris investigation software and expand globally, serving government and corporate clients in risk management. [UKTN]
ElevenLabs raised US$180m in Series C funding, tripling its valuation to US$3.3bn, as it expands its AI-powered text-to-speech technology that replicates human voices. [Tech.eu]
Final Round AI secured US$6.88m in seed funding to expand its AI-powered job search platform and launch an US$8m scholarship program for job seekers. [Feed the AI]
Gallabox has raised US$3.5m in funding to expand its no-code conversational AI platform that helps businesses automate WhatsApp interactions across 45 countries. [FinSMEs]
GetWhys has raised US$2.75m in seed funding to expand its AI-powered customer insights platform that helps teams gather competitive intelligence and develop marketing strategies. [FinSMEs]
HerculesAI has rebranded from ZERO, raised US$26m in Series B funding, and launched Verify, an AI-powered billing review platform. [LawSites]
Indigo secured US$8m in seed funding to develop its AI-powered real estate negotiation platform, streamlining transactions for agents, buyers, and sellers. [Feed the AI]
Ivo has secured US$16m in Series A funding to expand its AI-powered contract review solution, building upon its existing base of 150 corporate clients. [Artificial Lawyer]
Jump has raised US$20m in Series A funding to enhance its AI assistant service for financial advisors, focusing on workflow automation and enterprise compliance controls. [FinSMEs]
Krutrim has received a US$230m investment from founder Bhavish Aggarwal while open-sourcing its AI models and announcing plans for India’s largest supercomputer. [TechCrunch]
Lawpath has secured US$10m in funding and is launching enhanced AI legal services. [Artificial Lawyer]
Little Otter, a family mental health startup using AI technology, has raised US$9.5m in funding to support its digital healthcare platform. [Fortune]
Lorikeet has raised US$9m in funding to enhance its AI-powered customer service technology and expand enterprise capabilities. [FinSMEs]
Napier AI has secured investment to enhance its AI-powered financial crime detection platform and expand its global presence. [FinTech Global]
Neuralk-AI has secured US$4m in funding to transform how organisations harness tabular data for intelligent, actionable decisions. [Tech Funding News]
Omi secured US$2m in funding to develop an AI wearable device and brain-computer interface technology. [Omi]
Pandektes has raised €2.9M in seed funding to expand its AI-powered legal research platform that aims to unify and make searchable millions of EU court rulings. [Tech.eu]
Perspective AI secured US$4m in seed funding to expand its AI platform that decodes customer intent and transforms conversations into actionable business insights. [Feed the AI]
Presentations.ai has raised US$3m in seed funding after its AI-powered presentation platform attracted over 5m users and achieved profitability. [TechCrunch]
Prior Labs, an AI startup pioneering foundation models for spreadsheets and databases, has raised €9m in pre-seed funding. [Tech Funding News]
qeen.ai has raised US$10m in funding to expand its e-commerce AI platform that helps merchants automate tasks and boost sales through autonomous agents. [FinSMEs]
Savant Labs raised US$18.5m in Series A funding to expand its AI-powered analytics platform that helps business analysts manage data more efficiently. [FinSMEs]
Semeris has raised US$4.3m in funding to expand its AI-powered legal document analysis platform serving financial institutions and law firms. [FinSMEs]
Tana has raised US$14m in Series A funding for its AI-powered workspace platform that transforms unstructured information into structured knowledge. [Tech.eu]
TrueFoundry has raised US$19m in Series A funding led by Intel Capital to expand its AI deployment platform that helps enterprises manage and scale AI applications. [Business Wire]
Acquisitions
Data443, a data security and privacy software company, announced the acquisition of the IP and operational assets of Breezemail.ai, a provider of AI-powered email management technology. [EIN Presswire]
LavaReach has been acquired by Landbase, combining their AI-powered sales platforms to enhance go-to-market automation capabilities through Landbase’s GTM-1 Omni model. [PRWeb]
Litera has acquired Peppermint Technology, a UK-based provider of Microsoft-integrated cloud technology for law firms, expanding its legal technology portfolio. [LawSites]
Onit acquired Legal Files Software, a 30-year-old provider of legal case management software serving over 500 corporate, government, and law firm customers. [LawSites]
Smarsh has acquired CallCabinet, expanding its communications compliance platform with cloud-native call recording and AI-powered voice analytics capabilities for regulated organizations. [Business Wire]
Sorenson Communications announced the acquisition of Hand Talk and OmniBridge, two companies that specialize in sign language interpreting technology. [Slator]
XTM International has acquired Transifex, a SaaS-based localization and translation management platform. [MultiLingual]
There’s More
Doomers, a play by Matthew Gasda, dramatizes OpenAI's 2023 leadership crisis through a fictionalized account exploring AI safety concerns and Silicon Valley culture. [TechCrunch]
Emtherical has pioneered emotion-focused AI technology that understands human feelings, targeting emotional intelligence applications. [PRWeb]
Wiley has released ExplanAItions, a comprehensive study based on feedback from nearly 5,000 researchers worldwide examining AI usage patterns and adoption across academic research. [Business Wire]
The Back Cover
Here’s Clara Stegall with a cover of Colin Hay’s Waiting For My Real Life To Begin.
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!