This Week in NLP #338
Keep up with what happened in NLP in the week ending Friday 9th May 2025.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Apple has begun exploring AI-powered search engines for Safari, potentially threatening Google's US$20B annual deal as the browser’s default search provider. [CNN]
OpenAI has abandoned its nonprofit-to-profit conversion plan, instead restructuring its commercial arm as a public benefit corporation while maintaining nonprofit oversight. [The Verge]
And OpenAI is acquiring AI coding tool developer Windsurf for US$3B, marking its largest acquisition and expanding its presence in the developer tools market. [ITPro]
Microsoft has considered adding Musk’s Grok chatbot to Azure AI Foundry, potentially straining its relationship with OpenAI. [Yahoo Finance]
Mistral AI has launched Le Chat Enterprise, a privacy-focused AI assistant platform powered by its new Medium 3 model, targeting enterprise customers with cross-application support and data protection features. [VentureBeat]
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
Amazon's Alexa+ AI assistant has reached 100,000 US users, though CEO Andy Jassy acknowledged the technology remains primitive and lacks some promised features. [TechRadar]
Anthropic has launched a program offering up to $20,000 in API credits to researchers conducting high-impact scientific projects, particularly in biology and life sciences. [TechCrunch]
Apple has partnered with Anthropic to develop an AI-powered code assistance platform, integrating Claude Sonnet into an updated version of Xcode for internal use. [Verdict]
Meta has increased its projected 2025 infrastructure spending to US$72B, citing Trump-era tariffs and growing AI competition as key factors behind the US$7B jump. [The Register]
Nvidia has clashed with Anthropic over the AI lab’s support for stricter US export controls on AI chips to China, dismissing claims about creative smuggling methods. [Yahoo Finance]
OpenAI has announced deployment changes for ChatGPT following a problematic update that made the AI excessively agreeable, including new testing phases and enhanced safety reviews. [TechCrunch]
Here’s OpenAI’s blog post on the abandonment of its nonprofit-to-profit conversion plan. [OpenAI]
But Elon Musk has decided to pursue his lawsuit against OpenAI despite the company’s revised plan. [Reuters]
And Microsoft has withheld approval of OpenAI's restructuring plan, seeking to protect its US$13.75B investment in the AI company. [Bloomberg]
OpenAI has projected reducing Microsoft's revenue share from 20% to 10% by 2030, despite their extensive partnership and Microsoft’s multi-billion-dollar investment. [TechCrunch]
Salesforce CEO Marc Benioff has warned that Microsoft's alleged anti-competitive behavior toward Slack before its US$27B acquisition could be repeated with OpenAI. [ITPro]
ServiceNow has launched an AI-enhanced CRM platform to challenge Salesforce, integrating agentic AI for improved customer service and data management. [TechRadar]
Thinkagents.ai has launched an open-source framework enabling autonomous AI agents to operate across decentralized networks, challenging Big Tech’s control over AI. [Cointelegraph]
Sovereign AI
America’s tech dominance has eroded as China, London, Europe, Gulf States, and India have strengthened their AI capabilities and talent retention while US policies discouraged immigration. [Semafor]
Khazna, a Dubai-based data center firm, has begun exploring US investment opportunities following UAE’s US$1.4T commitment to the American economy. [Semafor]
Malaysia’s Science Ministry has adopted AI governance guidelines as an interim measure while working toward formal legislation, following concerns about AI-enabled scams. [Malay Mail]
Microsoft, OpenAI, and AMD executives have testified before a US Senate committee about maintaining America’s AI leadership against China’s growing capabilities. [Android Headlines]
And Microsoft, Etsy, Uber, and other major companies’ CEOs have signed an open letter advocating mandatory K-12 AI and computer science education in the United States. [TechCrunch]
Nvidia's CEO warned Congress that export restrictions on AI chips could enable Huawei's global expansion while threatening American technological leadership and the company’s market position. [Yahoo Finance]
OpenAI has announced plans to expand its US$500B Stargate AI project internationally, seeking to establish up to 10 facilities across multiple partner nations. [Yahoo Finance]
The Trump administration is planning to replace Biden’s three-tiered AI chip export control system with a simpler framework that would likely benefit Israel’s access to advanced semiconductors. [CTech]
The UAE has introduced AI into its public-school curriculum from kindergarten through 12th grade, aiming to strengthen its position as a regional AI leader. [Bloomberg]
Feature Creeps
Anthropic has expanded Claude’s capabilities with new Integrations for connecting to third-party apps and enhanced Research features that search across multiple data sources. [Anthropic]
Anthropic has launched web search capabilities for its Claude AI assistant, enabling developers to access current information through its API amid growing competition in AI-powered search. [VentureBeat]
And Anthropic's Claude is preparing to launch a push-to-talk voice mode feature in its mobile app, offering web search, file uploads, and multiple voices. [Analytics India Magazine]
Google's NotebookLM has expanded beyond desktop with mobile apps launching May 20, bringing its AI-powered note-taking and research features to Android and iOS devices. [TechCrunch]
Google Workspace has launched AI-powered enhancements across its suite, including Gemini’s Canvas workspace, audio summaries, and improved Slides features to boost business productivity. [Forbes]
Google has integrated Gemini into Sheets, enabling users to automate data analysis, create reports, and build dashboards directly within their spreadsheets. [ZDNet]
Google has announced supervised Gemini access for under-13s through parent-managed accounts, while acknowledging potential risks and implementing safety measures. [Engadget]
Google has launched a Gemini iPad app with split-view capability, expanding its AI assistant’s reach across Apple devices. [Engadget]
Google's Gemini has launched direct image editing capabilities and expanded file upload limits, allowing users to edit up to 10 images simultaneously using AI-powered tools. [TechRadar]
Google has launched automatic ‘implicit caching’ for its Gemini API, promising 75% cost savings on repetitive requests without requiring manual configuration from developers. [TechCrunch]
And Google has launched new AI-powered defenses in Chrome, including Gemini Nano for desktop scam protection and notification warnings for Android users. [TechCrunch]
Microsoft has unveiled AI-powered features for Windows 11 Copilot+ PCs, including a natural language settings assistant and improvements to Photos, Snipping Tool, and Narrator. [TechRadar]
OpenAI has rolled back ChatGPT-4o due to sycophancy issues and is considering implementing preset personalities as part of its solution to improve user interaction. [TechRadar]
Wix has launched its Model Context Protocol Server, enabling users to connect AI tools with Wix’s business functionality through natural language prompts. [GlobeNewswire]
Hype Bubble?
AWS has reported that one UK business is adopting AI every minute. [TechRadar]
A global study by DigitalRoute revealed that 71% of CFOs have struggled to monetize AI effectively, despite considering it mission-critical for their five-year strategy. [TechRadar]
IBM's survey of 2,000 CEOs has revealed that fear of missing out, rather than proven returns, is driving AI investments, with only 25% of initiatives meeting ROI expectations. [The Register]
And IBM has replaced hundreds of HR jobs with AI while increasing overall employment through strategic reallocation of resources to higher-value roles. [Yahoo Finance]
ZDNET-Aberdeen research has revealed that only 8% of Americans would pay extra for AI features, with 71% explicitly rejecting additional costs for AI capabilities. [ZDNet]
Hardware
Apple's silicon design team is developing chips for smart glasses, AI servers, and advanced Macs, with glasses processor production planned for late 2026 or 2027. [Yahoo Finance]
DigitalOcean has launched new GPU Droplets featuring Nvidia RTX 4000, RTX 6000, and L40S GPUs to support AI workloads and graphics processing. [Business Wire]
Huawei has developed the Ascend 910D AI chip to challenge Nvidia's H100, though analysts doubt it can match Nvidia’s software ecosystem and performance advantages. [TechRadar]
IBM has launched the LinuxONE Emperor 5 mainframe, featuring the Telum II processor, advanced AI capabilities, and enhanced security features for enterprise workloads. [ZDNet]
Imagination Technologies has unveiled its E-Series GPUs, combining graphics and AI processing capabilities with up to 200 TOPS performance for edge computing applications. [VentureBeat]
Nvidia's AI chips have become the focus of proposed US legislation aimed at tracking their locations after sale to prevent smuggling into China. [Yahoo Finance]
Nvidia has informed major Chinese clients it is redesigning AI chips to comply with US export rules while maintaining sales to the Chinese market. [DealStreetAsia]
And Nvidia has gained potential benefits from US consideration of easing AI chip export restrictions to the UAE. [Yahoo Finance]
The Trump administration’s plans to replace Biden’s AI chip export controls with simpler rules caused Nvidia's stock to rise 3%. [Ars Technica]
Seagate has announced plans to develop a 100-terabyte hard drive by 2030, tripling its current maximum capacity amid growing AI-driven data storage demands. [CNBC]
Consumer AI
eBay has launched an AI shopping assistant that helps buyers find items through chat conversations and image uploads, even when they’re unsure what they want. [ZDNet]
MeshTek has launched Talk To Your Lights, an AI application enabling natural voice control of home lighting through its long-range wireless protocol. [Speech Technology Magazine]
Netflix has launched an OpenAI-powered search feature allowing users to find content through conversational descriptions, initially rolling out in beta on iOS devices. [ZDNet]
It’s Only a Model
Google DeepMind has launched Gemini 2.5 Pro ‘I/O’ edition, surpassing competitors in coding benchmarks and offering enhanced web app generation capabilities from single prompts. [VentureBeat]
IBM CEO Arvind Krishna has championed smaller, domain-specific AI models as more efficient and cost-effective solutions for enterprises than large, general-purpose models. [ITPro]
And IBM has released Granite 4.0 Tiny Preview, a compact hybrid Mamba-2/Transformer language model that runs multiple long-context sessions on consumer GPUs. [IBM]
Mistral AI has launched Mistral Medium 3, an efficient AI model performing at 90% of Anthropic's Claude Sonnet while offering lower costs and multi-platform deployment options. [TechCrunch]
Nvidia has released Parakeet-TDT-0.6B-v2, an open-source speech recognition model that transcribes audio 60 times faster than real-time with near-commercial accuracy levels. [VentureBeat]
ServiceNow has partnered with Nvidia to develop Apriel Nemotron 15B, a new AI reasoning model offering improved performance and lower costs for enterprise applications. [Business Wire]
AI models’ training compute requirements have doubled approximately every five months since 2010, driven primarily by increased spending and hardware improvements. [Epoch AI]
Whose Data?
Google has maintained the ability to train its search-specific AI on web content that publishers opted out of using for other Google AI products. [Yahoo Finance]
IAB Tech Lab is developing an API framework to help publishers monetize and control their content usage by LLMs, replacing unstructured web scraping. [The Drum]
The LLM Ecosystem
Asana has launched Smart Workflow Gallery, offering prebuilt AI-powered workflows to help organizations scale AI usage and boost employee productivity across departments. [Business Wire]
Endava has partnered with Google Agentspace to implement autonomous AI systems that streamline enterprise workflows and boost organizational productivity. [Business Wire]
Epoch AI's study reveals data centers could cost US$200B to build by 2030, with private companies now owning 80% of AI supercomputing infrastructure. [ITPro]
Franz Inc has launched AllegroGraph 8.4, enhancing natural language query capabilities for enterprise-level neuro-symbolic AI solutions through improved knowledge graph technology. [EIN Presswire]
IBM has expanded its partnership with Oracle to integrate watsonx AI products into Oracle Cloud Infrastructure. [Yahoo Finance]
IBM has launched watsonx Orchestrate, an AI platform enabling businesses to deploy autonomous agents across enterprise applications while supporting both no-code and pro-code development. [Yahoo Finance]
And IBM has transferred three AI development tools - Docling, BeeAI, and Data Prep Kit - to the Linux Foundation’s LF AI & Data Foundation, strengthening open-source AI development. [BigDATAwire]
Lumen Technologies has partnered with IBM to develop edge-based AI solutions that integrate watsonx with Lumen’s infrastructure, helping businesses scale AI deployment more efficiently. [Yahoo Finance]
Microsoft has embraced Google's Agent2Agent protocol by adding support to Azure AI Foundry and Copilot Studio, enabling AI agents to communicate across different platforms. [TechCrunch]
OpenRouter has launched a unified API marketplace enabling developers to access multiple AI language models through a single interface, simplifying integration and management of various providers. [KDnuggets]
Peloton Consulting Group has launched an AI Framework for Oracle Cloud Fusion Applications, providing a structured seven-phase approach for enterprise-wide AI implementation. [EIN Presswire]
Rafay Systems has launched a Serverless Inference offering that helps GPU cloud providers deliver AI services while accelerating enterprise AI adoption. [Business Wire]
SAS has expanded its data capabilities by announcing Data Maker for synthetic data generation. [ITPro]
ServiceNow has launched AI Control Tower, a centralized platform for managing AI agents and workflows, alongside AI Agent Fabric for enhanced agent-to-agent communication. [Business Wire]
And ServiceNow has launched a new Workflow Data Network ecosystem and announced plans to acquire data.world to enhance AI-powered data integration and governance capabilities. [Business Wire]
Plus, ServiceNow has partnered with AWS to launch bi-directional data integration and automated workflow orchestration, enabling unified enterprise data management and real-time insights. [Business Wire]
SOAX has launched a redesigned Web Unblocker featuring enhanced anti-bot bypass capabilities and improved data pipeline functionality for AI and LLM applications. [EIN Presswire]
Agentic AI
AccelerateX has sought to replace 70,000 federal workers with AI agents. [Wired]
AgentSense has launched two AI agents, DocuCortex and DataGenie, alongside its website to help CPG, healthcare, and manufacturing businesses streamline operations and process data more efficiently. [Business Wire]
Algolia has launched real-time AI agents that integrate with Salesforce and Adobe platforms to provide context-aware responses using live, structured data across enterprise systems. [Business Wire]
Amnic has launched an AI-powered FinOps operating system that uses context-aware agents to automate and streamline cloud cost management tasks on AWS. [Business Wire]
Blue Yonder has launched five AI agents and a supply chain knowledge graph to transform supply chain management through faster decision-making. [Business Wire]
Hugging Face has launched Open Computer Agent, a free cloud-based AI that can perform basic computer tasks but suffers from slow performance and occasional errors. [TechCrunch]
Jeeva AI has emerged from stealth mode, launching an agentic AI platform that automates sales processes through autonomous lead discovery and multichannel engagement. [Business Wire]
Nutanix has launched an updated Enterprise AI solution featuring enhanced Nvidia AI Enterprise integration to accelerate enterprise-level Agentic AI application deployment. [GlobeNewswire]
OpenAna has launched Ana, claimed to be the world’s first autonomous AI/ML engineer, able to independently build and manage ML systems. [EIN Presswire]
Otter has launched an AI meeting assistant that answers spoken questions in real-time, schedules follow-ups, and integrates with company knowledge bases and CRM systems. [Fast Company]
ServiceNow has launched autonomous AI agents for security and risk management, partnering with Cisco and Microsoft to help enterprises shift toward autonomous defense capabilities. [Business Wire]
SoundHound has launched Amelia 7.0, an AI platform combining voice technology and autonomous agents to handle complex business tasks through natural interactions. [Business Wire]
Other LLM Sightings
Amazon has launched ‘Enhance My Listing’, an AI-powered tool that helps merchants automatically improve their product listings with suggested titles, descriptions, and missing details. [TechCrunch]
Bigeye has launched bigAI, an AI solution that helps enterprise data teams resolve issues faster through real-time, actionable incident summaries. [EIN Presswire]
Darwin has launched a free AI Policy Wizard to help government agencies create customized AI governance policies quickly and efficiently. [PRWeb]
DataRobot has launched a federal AI application suite designed to help government agencies implement secure, mission-critical AI solutions in high-security environments. [Business Wire]
Diald AI has launched an AI-powered investment memo service that analyzes property data from 1.7 million sources, saving real estate professionals time and money. [Business Wire]
Electron has launched an unofficial Clippy-themed interface for running local AI language models across Windows, macOS, and Linux platforms. [The Register]
FutureHouse has launched an AI platform featuring four specialized agents that help scientists accelerate research by analyzing literature and planning experiments more effectively than humans. [siliconANGLE]
Groupe Sentinella has integrated Fiona AI technology to streamline candidate interviews and reduce hiring times across its Canadian recruitment operations. [EIN Presswire]
Highspot has launched its Spring '25 release, introducing AI-powered features for enhanced digital rooms, content workflows, and buyer engagement to improve sales effectiveness. [Business Wire]
Klippa has launched enhanced AI-powered fraud detection within its DocHorizon platform to help businesses combat AI-driven document forgery. [Silicon Canals]
Korl has launched a multi-agent AI platform that combines data from various business systems to generate customized customer communications and presentations. [VentureBeat]
Particle has expanded its AI-powered news aggregation service to the web through Particle.news, offering AI summaries, entity pages, and publisher links alongside its existing mobile app. [TechCrunch]
Scalafai has launched Safia, an AI-powered productivity platform that integrates with 300+ enterprise sources to transform scattered data into actionable team insights. [EIN Presswire]
ServiceNow has launched Core Business Suite, an AI-powered solution that transforms core business processes by connecting employees, suppliers, systems, and data in one platform. [Business Wire]
Trullion has launched Trulli, an AI assistant specifically trained for accounting workflows, designed to help finance professionals automate tasks and analyze documents. [Business Wire]
Risks and Responses
Being Human With AI has launched a free online curriculum helping children aged 9-13 navigate AI risks through interactive lessons focused on ethical discernment. [PRWeb]
Cohere's research head has criticized LM Arena's chatbot leaderboard for allegedly allowing major AI companies to manipulate rankings through preferential testing practices. [BetaKit]
Enkrypt AI has discovered that Mistral AI's multimodal AI models were significantly more vulnerable to generating harmful content through image-based prompt injections than competing systems. [ZDNet]
The European Broadcasting Union has led a global media coalition in launching five steps to ensure AI technology supports fact-based news and fights misinformation. [Editor & Publisher Magazine]
EY's global survey revealed Americans trust AI’s accuracy but rarely review its output, with only 24% checking AI-generated content compared to 31% worldwide. [The Decoder]
Giskard has discovered that instructing AI chatbots to provide concise answers increased their tendency to generate false information, particularly when addressing ambiguous topics. [TechCrunch]
Google's Gemini 2.5 Flash has scored lower on safety tests than its predecessor, showing increased likelihood to generate policy-violating content. [TechCrunch]
Google has deployed AI-enhanced security systems across its products to combat scams, including improved detection in Search, Chrome, and Android messaging apps. [ZDNet]
Instagram co-founder Kevin Systrom has criticized AI companies for prioritizing user engagement through excessive follow-up questions rather than focusing on providing quality answers. [TechCrunch]
Meta is developing smart glasses with always-on AI capabilities for facial recognition and reminders, despite potential privacy concerns and battery life limitations. [TechRadar]
OpenAI's latest models, GPT o3 and o4-mini, have demonstrated increased hallucination rates compared to earlier versions, with errors occurring in up to 79% of responses. [TechRadar]
And OpenAI has confirmed a bug in ChatGPT that allowed minors to access and be encouraged to request explicit sexual content. [TechCrunch]
OpenAI has rolled back an April 25th GPT-4 update that made the model overly sycophantic, explaining the causes and implementing new safeguards against similar issues. [OpenAI]
Open-source AI models have demonstrated significant gender bias in hiring recommendations, favouring men for high-paying roles while relegating women to lower-wage positions. [The Register]
Oxford researchers have found that people using AI chatbots for medical advice made poorer health-related decisions than those using traditional information-seeking methods. [TechCrunch]
Singapore has released a blueprint for global AI safety collaboration, bringing together researchers from the US, China, and Europe to prioritize research and risk management. [Wired]
Trend Micro has identified critical security vulnerabilities in Nvidia Riva’s API endpoints, prompting recommendations for enhanced security measures in deployments of the AI speech service. [Slator]
This summer’s World Science Fiction Convention organizers have apologized for using ChatGPT to screen potential panelists at their Seattle convention, following backlash from science fiction authors and resignations from staff. [The Register]
Conversational AI
Airbnb has launched an AI customer service bot in the US. [TechCrunch]
CallMiner has enhanced its AI Assist platform with new supervisor and research agents to help organizations better analyze and act on customer conversation data. [Business Wire]
Cerence AI has partnered with Code Factory to launch VoiceTopping, bringing conversational AI technology to self-service kiosks across multiple industries. [GlobeNewswire]
Eventdex has launched an AI-powered voice and chat assistant within its event management software to provide real-time support for attendees and organizers. [EIN Presswire]
GIDR.ai has partnered with ServiceNow to launch a multimodal conversational AI platform that enhances service productivity and regulatory compliance across industries. [EIN Presswire]
Kore.ai has formed a strategic partnership with Microsoft, integrating its AI platform with Microsoft’s cloud ecosystem to accelerate enterprise-wide AI transformation. [Business Wire]
Lowe's has deployed Mylow Companion, an OpenAI-powered voice assistant that helps store associates access product information and provide expert advice across 1,700+ locations. [Speech Technology Magazine]
Mem0 has launched two memory architectures that help AI language models maintain coherent conversations by dynamically managing information from extended interactions. [VentureBeat]
NICE has partnered with ServiceNow to create an AI-powered customer service automation platform that connects front-office interactions with enterprise workflow management. [Business Wire]
RateGain has launched UNO VIVA, an AI-powered voice agent that manages hotel reservations in multiple languages while providing 24/7 automated customer service. [Speech Technology Magazine]
Reactor Data has launched Electron AI, an embedded conversational assistant helping data teams create mappings, transformations, and pipelines for major data platforms. [Business Wire]
RingCentral has launched RingCX for Salesforce Service Cloud Voice, integrating digital and voice communications within Salesforce’s platform to enhance customer service delivery. [Business Wire]
Rio Moving has launched ‘Jane’, an AI phone assistant trained on 30,000 customer calls to handle reservations and inquiries around the clock. [EIN Presswire]
ServiceNow has unveiled an AI-powered CRM platform that unifies sales, fulfillment, and service functions to deliver end-to-end customer experiences on a single system. [Business Wire]
SourceNow has launched Sage AI, a multilingual AI agent embedded within its VMS platform to enhance workforce management and automation capabilities. [Business Wire]
Talkdesk has launched its Utilities Experience Cloud platform, combining AI-powered customer service with utility-specific workflows to modernize how providers handle customer interactions. [GlobeNewswire]
And Talkdesk has also launched Express, an AI-powered customer service platform designed to give small businesses enterprise-grade capabilities at an affordable price with simplified setup. [GlobeNewswire]
Be Real
The Academy of Motion Picture Arts and Sciences has established that AI use in filmmaking won’t disqualify Oscar nominations. [eWeek]
Arizona courts have made history by allowing an AI-generated version of murder victim Christopher Pelkey to deliver an impact statement during his killer’s sentencing hearing. [ABC15 Arizona]
GOAT Gaming has launched Amy, an AI gaming agent that plays against users, manages social interactions and automated earnings, in a preview of the company’s upcoming AlphaGOAT Pros platform. [VentureBeat]
Reddit has announced plans to implement third-party user verification services to combat AI bots. [TechCrunch]
Voice News
Rime has launched Arcana, a text-to-speech model that generates ultra-realistic voices with natural speech patterns and laughter. [Rime]
NBC has partnered with Jim Fagan’s family to resurrect the late narrator’s voice through AI for its upcoming NBA coverage starting in October. [The Verge]
Nvidia has released Parakeet TDT 0.6B, an open-source transcription tool with industry-leading accuracy that can process 60 minutes of English audio per second. [eWeek]
SYNQ Technology has launched AI-powered voice solutions that transform traditional retail radio communications into intelligent, data-driven systems using Microsoft AI services. [EIN Presswire]
Document AI
Appian has launched enhanced AI platform features, including Agent Studio and AI Document Center, to help organizations build and scale intelligent process applications more effectively. [Appian]
Crawford Technologies has launched AccessibilityNow Gateway Enterprise, a cloud-first solution that converts corporate documents into accessible formats for visually impaired readers. [Crawford Technologies]
Insurants has launched a Submission Processing solution that automates the extraction and summarization of data from insurance-related emails and documents to streamline underwriting processes. [Insurants]
IntelliDocX has launched an AI Content Service for Azure that transforms SAP business content into intelligent assets using Microsoft Azure AI capabilities. [PRWeb]
InterScan has launched JetStream AI 5.3, an AI-powered document management platform that automates complex document processing and data extraction for enterprises. [EIN Presswire]
Oxford’s Bodleian Libraries has revealed the title and author of a 2,000-year-old Herculaneum scroll using AI-powered virtual unwrapping techniques. [Gizmodo]
Prophia has launched an AI-powered instant lease abstraction tool that processes commercial real estate documents faster and more affordably than traditional methods. [Business Wire]
TCG Process has launched OCTO, an AI-driven automation platform that orchestrates workflows and integrates with its existing DocProStar IDP solution. [Intelligent Document Processing Community]
Vanilla has launched enhanced AI capabilities for estate planning, featuring document analysis tools and automated summaries to help advisors serve clients more efficiently. [Business Wire]
Translation
Baidu has filed a patent for an AI-powered system designed to translate animal vocalizations and behaviors into human language, focusing initially on cat communication. [Yahoo Finance]
PubHive has launched AI-powered translation capabilities to help life sciences teams streamline global literature reviews and regulatory compliance across multiple languages. [EIN Presswire]
Silence Speaks has launched an AI-powered avatar system that translates text into sign language, aiming to help millions of deaf people overcome communication barriers globally. [Wired]
Search
Alibaba has developed ZeroSearch, a system that trains AI to perform searches without using real search engines, reducing costs by 88% while matching Google's performance. [VentureBeat]
ArcBlock has launched ArcSphere, an AI-native browser featuring intelligent assistance for iOS and Android devices. [EIN Presswire]
BrandRank.AI has launched a SaaS dashboard that helps marketers measure and optimize their brand visibility and trust across AI-powered search engines. [EIN Presswire]
Google's search volume declined in Safari for the first time in 22 years, prompting Apple's Eddy Cue to attribute the drop to AI’s growing influence. [The Verge]
HuskyTail Digital has launched AISO (AI Search Optimization), helping businesses optimize their content for AI platforms like ChatGPT, Perplexity, and Gemini. [EIN Presswire]
LinkedIn has launched AI-powered job search tools, including natural language search capabilities, personalized coaching, and improved job-matching insights for its users. [ZDNet]
OpenSearch has achieved a 9.5x performance boost in its 3.0 release through GPU-powered indexing, particularly benefiting vector database workloads with a 9.3x speed increase. [BigDATAwire]
Pinterest has enhanced its visual search capabilities with AI-powered features that help users better refine searches and identify specific aspects of images they like. [TechCrunch]
Reddit has integrated its AI-powered Reddit Answers feature into the main search bar, making it easier to find curated, experience-based responses from community discussions. [ZDNet]
Searchlight Cyber has added AI summarization and sentiment analysis capabilities to its Cerberus platform, helping investigators analyze dark web content more efficiently. [Business Wire]
Wiley has partnered with AWS to develop an AI agent for searching scientific literature, becoming the first publisher to launch such a tool on AWS’s platform. [Business Wire]
And Wiley has partnered with Perplexity to integrate its educational content into AI search capabilities, becoming Perplexity’s first education partner for trusted scholarly resources. [Business Wire]
Writing Assistance
Cornell researchers discovered that AI writing assistants homogenized language between US and Indian users, raising concerns about cultural erasure and Western bias. [eWeek]
Google has launched an iOS-exclusive ‘Simplify’ feature that uses dual Gemini AI models to make complex text more readable, showing improved comprehension in testing. [The Register]
Health Tech
CharmHealth has launched CharmEasyCare, a streamlined appointment booking system that matches patients with healthcare providers. [Business Wire]
EUCALIA has partnered with Hippocratic AI to develop Japan’s first non-diagnostic, patient-facing Japanese-language generative AI healthcare agent. [Business Wire]
Lightning Step's AI assistant LIA has become behavioral healthcare’s first ISO 42001-certified AI solution, reducing clinical documentation time while maintaining ethical standards and data privacy. [Business Wire]
Neurotone AI has partnered with Northeastern University to integrate its AI-powered auditory training technology into speech pathology and audiology programs. [PRWeb]
NewVue.ai has partnered with MD.ai to integrate AI-powered radiology reporting directly into its EmpowerSuite Radiologist Cockpit, streamlining workflow for radiologists. [EIN Presswire]
Pathway's medical AI model has achieved 96% accuracy on US Medical Licensing Examination questions, outperforming competitors like GPT-4 and Med-Gemini. [Business Wire]
RevenueWell has partnered with Mila Health to introduce an AI assistant that automates dental practice management tasks and provides 24/7 patient service. [Business Wire]
Scriptor Software has launched rScriptor, an AI-powered radiology reporting tool that decouples transcription from reporting. [Speech Technology Magazine]
UnitedHealth Group has deployed 1,000 AI applications across its healthcare operations, including claims processing, transcription, and software development. [The Wall Street Journal]
WELLSTAR Technologies has launched Nexus AI, a Canadian-built platform offering AI-powered medical documentation and clinical support tools to healthcare providers across Canada. [Business Wire]
Legal Tech
Bryter has launched Beamon AI, a comprehensive suite of AI agents for lawyers that combines workflow automation with AI capabilities for end-to-end legal processes. [Artificial Lawyer]
Claim Time Solicitors has implemented generative AI technology to detect fraudulent personal injury claims more quickly and accurately while supporting legitimate claimants. [PRWeb]
CobbleStone Software has partnered with Secure Operating Solutions to expand its CLM platform’s reach and enhance contract management services for customers. [PRWeb]
FasterOutcomes has partnered with Rob Levine Legal Solutions to integrate AI-powered legal automation tools and medical record retrieval services for law firms nationwide. [PRWeb]
Legal Decoder has launched LiSA, an AI-powered data analyst that helps legal departments analyze billing data through natural language conversations. [EIN Presswire]
LegalSifter has relaunched with new generative AI capabilities and expanded CLM features following its acquisition of Contract Logix. [Artificial Lawyer]
Nextpoint has partnered with Remote Legal to integrate their deposition software platforms, creating a comprehensive solution for legal professionals managing depositions. [PRWeb]
Practice AI has launched an integrated AI platform for law firms that automates document processing and demand letter generation while maintaining attorney oversight. [GlobeNewswire]
Vals AI has issued an open call for legal tech vendors to join its benchmarking studies, following its earlier evaluation of legal AI tools in February. [LawSites]
Ed Tech
Merlyn Mind has launched Ireland’s largest AI education pilot program, introducing voice-controlled teaching assistants in 48 schools. [The Clare Champion]
Panopto has partnered with Ukraine’s Ministry of Education to provide AI-powered video learning tools and conduct research across Ukrainian educational institutions. [PRWeb]
Funding
Anysphere, the maker of AI-powered coding tool Cursor, has secured US$900m in funding, reaching a US$9 billion valuation. [TechCrunch]
Carta Healthcare has secured US$18.25m in Series B1 financing to expand its AI-powered clinical data abstraction platform and grow its customer base, particularly in life sciences. [MobiHealthNews]
Databricks and KPMG have invested in LlamaIndex to enhance AI’s ability to process unstructured enterprise data through products like LlamaCloud and LlamaParse. [BigDATAwire]
Decagon is raising US$100m from Andreessen Horowitz and Accel at a US$1.5 billion valuation to expand its AI-powered customer service platform. [Forbes]
DecisionRules has secured €1.6m in funding to expand its cloud-native business rules platform that helps companies automate decision-making without complex technical requirements. [Tech.eu]
Diio has raised US$2.5m in seed funding to expand its AI sales assistant platform, which already serves over 200 Latin American companies. [FinSMEs]
Doubleword has secured US$12m in Series A funding to expand its self-hosted AI inference platform that helps enterprises deploy and manage AI models efficiently. [Silicon Canals]
Fastino has raised US$17.5m in seed funding for its small, task-specific AI models that run on affordable GPUs and deliver rapid enterprise solutions. [TechCrunch]
FirmPilot has raised US$11.7m in total funding for its AI legal marketing platform. [LawSites]
Fractile has announced plans to launch AI chips in 2027 that it claims will process LLMs more efficiently than competitors like Nvidia. [Tech.eu]
Hostie has raised US$4m in seed funding to expand its AI-powered customer communication platform for restaurants. [FinSMEs]
Justice Bid has secured US$4m in strategic investment from OnDean Forward to enhance its AI-powered platform for transparent outside counsel selection. [GlobeNewswire]
Legora, a Stockholm-based legal AI startup founded in 2023, has entered talks to raise US$85m from investors at a US$675m valuation. [Forbes]
LlamaIndex has received undisclosed investments from Databricks and KPMG to advance its document workflow platform and flagship products LlamaCloud and LlamaParse. [FinSMEs]
Luzia has raised US$13.5m in funding to expand its AI personal assistant service across Latin America, adding to its existing user base of 65 million people worldwide. [FinSMEs]
Nia, a London-based AI platform founded by 18-year-old Arlan Rakhmetzhanov, has secured $850K in pre-seed funding to develop autonomous engineering agents. [Silicon Canals]
Parloa has secured US$120m in Series C funding to expand its AI agent management platform across North America and Europe. [Silicon Canals]
StackOne has raised US$20m in Series A funding led by GV to advance its AI platform that connects B2B SaaS tools through 200+ connectors. [FinSMEs]
Thinkpilot has launched an AI-powered product management workspace with €600,000 pre-seed funding, integrating specialized agents and existing tools to support product teams’ decision-making. [Tech.eu]
Bezos Expeditions has led a $72-million investment in AI data solutions company Toloka, aiming to accelerate its growth and development of AI-human collaboration products. [Yahoo Finance]
Unblocked has raised US$20m in Series A funding to expand its AI-powered code assistant that helps developers understand and navigate complex codebases. [TechCrunch]
VSORA has secured US$46m in funding to develop Jotunn8, a European-made AI inference chip challenging Nvidia's market dominance. [TechRadar]
WisdomAI has launched an AI-powered data analytics platform that processes clean and unstructured data, securing US$23m in seed funding. [TechCrunch]
Acquisitions
CoreWeave has acquired AI developer platform Weights & Biases to enhance its cloud computing capabilities and accelerate AI development services. [FinSMEs]
Elastic has acquired Keep, an Israeli AI startup that developed an open-source platform to help tech teams manage and streamline alert systems and incident responses. [CTech]
PERSUIT has acquired Apperio, combining their outside counsel management platform with Apperio’s spend-management capabilities to enhance legal department operations and transparency. [LawSites]
Quansight has acquired Cobalt Speech and Language, integrating the latter’s multilingual voice-based AI systems and expertise into its open-source technology portfolio. [Speech Technology Magazine]
Weave has acquired TrueLark, an AI-powered virtual receptionist platform, for US$35m in cash and equity to enhance its healthcare practice software offerings. [FinSMEs]
XTM International has acquired TXTOmedia, combining their localization technology with AI-powered video automation to provide end-to-end multilingual content management solutions. [MultiLingual]
There’s More
An AWS study reveals that generative AI has surpassed cybersecurity as IT leaders’ top budget priority, with 45% of organizations prioritizing AI spending over traditional investments. [VentureBeat]
Neuralink's brain implant recipient Bradford Smith has enhanced his communication abilities by using Elon Musk’s Grok AI chatbot to help compose responses on social media. [MIT Technology Review]
xAI's chatbot Grok has frustrated MAGA supporters by providing factual, politically neutral responses despite being initially designed to appeal to conservative users. [Gizmodo]
The Back Cover
Here’s John Martyn with a cover of Bob Dylan’s Don’t Think Twice.
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!