IBL News

Category: Top News

  • AI-Created Personalized Reports with Visuals and Graphics Help Students to Improve Test Scores

    AI-Created Personalized Reports with Visuals and Graphics Help Students to Improve Test Scores

    IBL News | New York

    Researchers at the South China Normal University in Guangzhou, highly focused on teacher education and training, explored how AI-driven visual reports can transform assessment in K12. A 13-week study analyzed the performance of half of the students who received personalized reports with clear visuals, personalized feedback, and actionable insights versus the other half who received oral feedback from instructors.

    The “AI reports” group demonstrated a 12.8% improvement in test scores over those only receiving oral.

    In addition, the vast majority expressed interest in using the report for other classes.

    Experts highlighted that AI tools get learners more efficient and faster personalized feedback and actionable feedback.

    Students can see their learning journey in these reports and act on it. They review areas of strength, identify learning gaps, and plan where to focus their effort.

  • Google’s White Paper Explains How GenAI Is Building the Campus of Tomorrow

    Google’s White Paper Explains How GenAI Is Building the Campus of Tomorrow

    IBL News | New York

    More than half of colleges and universities are focused on staying ahead in the generative AI race by harnessing their data.

    A white paper from Google Cloud, released by the Chronicle of Higher Education, highlights that campuses are demonstrating remarkable resilience and creativity by using GenAI and multimodal models for personalized learning and other uses.

    “With the right approach and a trusted platform, GenAI can help your institution personalize content across every medium, gain valuable insights from student data, and achieve measurable ROI —all with enterprise-grade security and scalability,” says the report.

    Roy Daiany, Industry Director for Education and Careers at Google, explained, “A university’s first-party data is its competitive differentiator. By effectively harnessing that data and combining it with AI, institutions are able to lower costs and increase efficiency.”

    • Download the Report (PDF)

    [Disclosure: ibl.ai, the parent company of iblnews.org, is a partner of Google Cloud]

  • OpenAI Launched ‘Tasks’, a Basic Agent that Sets Reminders or Recurring Tasks

    OpenAI Launched ‘Tasks’, a Basic Agent that Sets Reminders or Recurring Tasks

    IBL News | New York

    OpenAI rolled out a feature named “Tasks” this week, which allows the AI assistant to do reminders or recurring tasks by following up with a push notification on whatever platform is enabled.

    This new feature, in beta, is for paying users of ChatGPT Plus, Team, and Pro around the globe this week.

    Examples of tasks include “Every Friday, give me a weekend plan based on my location and the weather forecast” or “Give me a news briefing every day at 7 a.m.”

    The new task feature appears to be OpenAI’s first step into AI agents or advanced agentic systems that can act somewhat independently.

    Users can access tasks by selecting “4o with scheduled tasks” from a dropdown menu in ChatGPT.

     

  • ACI Introduced Its Marketplace with 18,000 Training Courses and AI-Mentorship

    ACI Introduced Its Marketplace with 18,000 Training Courses and AI-Mentorship

    IBL News | New York

    Colorado Springs-based ACI Learning announced the addition of a marketplace functionality to its workforce development and training platform.

    Along with this feature, the company introduced 18,000 new training modules, allowing organizations to customize their learning ecosystems.

    The new myACI Marketplace offers access to new content in application development, leadership training, and productivity, thanks to partnerships with BizLibrary, Clip Training, Scrimba, Net Ninja, Bigger Brains, and other providers.

    • BizLibrary – Delivering expertise in business operations and human resources training with over 12,500 courses spanning Business Skills, HR Compliance, Software, Leadership & Management, Workplace Safety, Sales & Service, and IT.
    • Clip Training – Providing 5,700 courses focused on Cybersecurity Training, Phishing Training & Testing, and Microsoft 365 User Training.
    • Scrimba – Specializing in technical development with 60 courses in JavaScript, Python, React, TypeScript, AI, HTML, CSS, Tailwind, Imba, and UI Design.
    • Net Ninja – Will add more than 80 courses to myACI Marketplace that cover various programming languages, frameworks, and tools, including JavaScript, HTML, CSS, React, Node.js, Vue.js, Django, and more.
    • Bigger Brains – Providing 259 courses to enhance professional skills across various fields, including training in accounting, finance, business, communication, cybersecurity, data analysis, DEI, Google Workspace, HIPAA compliance, management, and productivity tools like Microsoft Office and Excel.

    “With myACI Marketplace, businesses can now build tailored learning experiences that are fully scalable and adaptable to their needs,” said Scott Horn, CEO at ACI Learning.

    The myACI Marketplace has a custom management system with AI-driven mentorship tools to personalize learning experiences.

    The workforce development company plans to add new content, partners, and AI-enhanced tools continuously.

    [Disclosure: This AI platform has been powered by ibl.ai, the parent company of iblnews.org]

  • An Open Source High Performance Video Generation Model from a Chinese Company

    An Open Source High Performance Video Generation Model from a Chinese Company

    IBL News | New York

    The Chinese company Tencent presented a framework for a large video generation model called “HunyuanVideo.”

    This Text-to-Video model was released as open source with its code and weights.

    HunyuanVideo was trained with over 13 billion parameters, making it the largest open-source model.

    “It exhibits performance in video generation that is comparable to, if not superior to, leading closed-source models,” said the authors.

    “It outperforms state-of-the-art models, including Runway Gen-3, Luma 1.6, and 3 top performing Chinese video generative models.”

  • Google Gemini’s ‘LearnLM’ Outperformed Other AI Models on Adaptive Learning

    Google Gemini’s ‘LearnLM’ Outperformed Other AI Models on Adaptive Learning

    IBL News | New York

    Google published a technical report this December stating that Gemini’s LearnLM outperformed other leading AI models when it comes to adhering to learning science principles.

    This includes explaining concepts appropriately, providing effective hints and guidance, interactively guiding learners, and encouraging active engagement.

    For example, when prompted to teach complex topics, LearnLM breaks content into chunks and uses guiding questions to encourage research instead of providing direct answers.

    “LearnLM is better able to follow instructions, support learning goals, and adapt to learner’s abilities, with pedagogy experts and testers preferring results from LearnLM over other models,” said Google’s report.

    LearnLM is Google’s AI models and capabilities that infuse learning science into Gemini, YouTube, and Classroom.

    This tool for personalized learning is available for preview access as an experimental model in Google AI Studio. Google’s goal is to add more improvements to Gemini models.

    Experts Libby Hills from the Jacobs Foundation and Owen Henkel from the University of Oxford discussed in a podcast how and why LearnLM was developed and what to expect next.

    • LearnLM API

  • Getty Images Plans to Merge with Rival Shutterstock

    Getty Images Plans to Merge with Rival Shutterstock

    IBL News | New York

    Getty Images said it would merge with rival Shutterstock on Tuesday to create a $3.7 billion visual licensed content company. The agreement will be subject to both companies’ antitrust regulatory and shareholder approval.

    The combined company would be named Getty Images Holdings, Inc and continue trading on the New York Stock Exchange under the “GETY” symbol. Getty CEO Craig Peters would remain chief executive.

    The deal underscores the demand for images and content from creators and platforms at a time when the growing use of generative AI tools such as Midjourney poses a threat to the industry.

    The two companies said the expected annual cost synergies would be between $150 million and $200 million by year three.

  • Nvidia Launched an Open-Source Video Model to Generate Synthetic Data and a Personal Supercomputer for AI

    Nvidia Launched an Open-Source Video Model to Generate Synthetic Data and a Personal Supercomputer for AI

    IBL News | New York

    Nvidia launched Cosmos this week, an open-source video model to generate synthetic data at scale, addressing the data scarcity in robotics.

    Trained on 20 million hours of video data and 9,000 trillion tokens, Nvidia Cosmos provides tools for creating physics-aware video simulations for various applications, mainly robotics and autonomous vehicles.

    It contains Difusion-based and Autoregressive model architectures, which, respectively, generate continuous tokens to create controllable visual simulations and use discrete tokens to predict future video frames.

    It supports text-to-video and video-to-video generation.

    Cosmos is available on Hugging Face, licensed under Apache 2, while preview access is offered at build.nvidia.com.

    Personal Supercomputer for AI

    On the other hand, Nvidia also introduced a personal AI supercomputer, DIGITS, for rapid prototyping of open-source Llama models.

    With the New GB10 Superchip, DIGITS debuts as the world’s most miniature AI supercomputer capable of running 200B-parameter models.

    It will be available from Nvidia and top partners in May, starting at $3,000.

  • Nvidia Partnered with CrewAI to Harness the Power of Agents

    Nvidia Partnered with CrewAI to Harness the Power of Agents

    IBL News | New York

    Nvidia announced a partnership with CrewAI yesterday to provide an agent orchestration solution with enterprise-grade performance and complete infrastructure control.

    Currently, PwC’s enterprise and startup Parloa are using this service.

    CrewAI now powers 4.5M+ Crews with multiple AI Agents weekly across industries.

    “The integration of CrewAI with NVIDIA AI Enterprise and NIM microservices marks a leap forward in production-grade AI agents,” said Justin Boitano, VP of Enterprise AI at NVIDIA

    The two companies see this transformation across back-office automation, sales & marketing operations, financial analysis, and coding and software automation.

    This production-ready demo showcased AI agents powered by Meta’s Llama 3.3 model on NVIDIA NIMs.

  • ElevenLabs Launched an Audio Tool for Agents with Law Latency

    ElevenLabs Launched an Audio Tool for Agents with Law Latency

    IBL News | New York

    The text-to-speech platform ElevenLabs launched a powerful AI tool named “Conversational AI” this month that allows to build agents with voice.

    The voice interaction shows low latency and interruption handling for back-and-forth conversation.

    The model allows for swapping between Gemini, Claude, and ChatGPT or bringing a custom implementation.

    It features external function calling and integrates with any third-party app to get real-time information.

    The agents are powered by 70 realistic-sounding AI voices in up to 32 languages and are built on users’ knowledge base.

    As use cases, ElevenLabs mentioned the following:

    • Customer Support to handle customer inquiries 24/7.

    • Scheduling for appointment booking and management, from healthcare to personal services.

    • Outbound Sales for cold calling at scale and qualifying leads.

    • Gaming experiences.

    • Education for personalized learning experiences, offering explanations, quizzes, and adaptive lessons based on a student’s progress.

    • Fan engagement with interactive experiences from trivia to personalized content recommendations.