- THE AI WAY
- Posts
- AI Transforms Global Education: MIT Unveils Brain-Like Data Processing and AI Wearables Take New Direction
AI Transforms Global Education: MIT Unveils Brain-Like Data Processing and AI Wearables Take New Direction
Niel de Kock
February 20, 2025
Welcome to the latest news in AI and AI in educational news.
Education AI News
Higher Education Embraces AI, but Strategic Implementation Lags
A new Educause 2025 AI Landscape Study reveals that 57% of educators now consider AI a strategic priority, up from 49% last year. However, only 22% of institutions have implemented institution-wide AI strategies, while most are rolling out AI on an ad-hoc basis across various departments.
Despite increasing interest, financial concerns remain a barrier. 34% of educators believe leadership is underestimating AI costs, and only 2% report new funding sources for AI projects. The top AI use cases in education include academic integrity, coursework, assessments, and curriculum design.
Notably, small and large institutions show similar trends in AI adoption, personal use, and future optimism, indicating that AI’s role in education is expanding across all institutional sizes.
MIT Study Finds Large Language Models Process Data Like the Human Brain
A new MIT study reveals that large language models (LLMs) process diverse data types using a centralized reasoning mechanism, similar to the human brain’s semantic hub in the anterior temporal lobe. The research, to be presented at the International Conference on Learning Representations, demonstrates that LLMs convert various data modalities—languages, images, audio, and code—into abstract, shared representations, using their dominant language as an intermediary.
Key findings include:
Modality Integration: LLMs represent different data types based on their underlying meaning, akin to how the brain integrates sensory inputs.
Dominant Language Processing: An English-trained model internally processes other languages, math, or code in English before generating outputs.
Predictable Intervention: Researchers successfully manipulated model outputs by intervening in its semantic hub with English text, even when the model handled different languages.
This insight could lead to more efficient multilingual and multimodal AI models, reducing redundancy in training while preserving language-specific nuances. The findings bridge AI and neuroscience, offering a deeper understanding of how advanced models process and reason about diverse data.
MIT Spinout ReviveMed Uses AI to Map the Body’s Metabolites and Uncover Disease Drivers
MIT spinout ReviveMed is revolutionizing disease research by leveraging AI-powered metabolomics to analyze how small molecules—such as lipids, cholesterol, and sugars—influence health and treatment outcomes. Founded by Leila Pirhaji PhD ’16 and Professor Ernest Fraenkel, the platform enables pharmaceutical companies and researchers to gain unprecedented insights into metabolic mechanisms, helping match patients with the most effective treatments.
Key innovations include:
AI-powered metabolic mapping: ReviveMed’s platform processes massive metabolomic datasets to uncover hidden biomarkers linked to cancer, Alzheimer’s, and cardiovascular conditions.
Patient-specific drug response analysis: ReviveMed helps pharmaceutical companies identify which patients will respond best to treatments by studying metabolic variations, potentially accelerating clinical trials and improving precision medicine.
Generative AI for metabolomics: The company has developed digital twins of patients based on 20,000 blood samples, creating AI models to simulate metabolic responses and improve drug development.
With partnerships with top global pharmaceutical companies, ReviveMed is democratizing metabolomics by making its AI models available to nonprofit researchers. It aims to bridge the gap between untapped metabolic data and effective therapies.
(https://news.mit.edu/2025/mit-spinout-maps-body-metabolites-uncovering-hidden-disease-drivers-0219)
Temple University Professor Uses AI and VR to Revolutionize Teacher Education
Temple University’s Di Liu won the 2025 Best Practice Award for the Innovative Use of Technology for revolutionizing teacher education with AI and Virtual Reality (VR).
VR-Powered Teaching – Uses immersive global classroom simulations to enhance cultural awareness and teaching skills.
AI for Multilingual Learning – Integrates HeyGen & ChatGPT to translate course materials into 26 languages, breaking language barriers.
Inclusive Education – Prepares future teachers to support diverse, multilingual learners using cutting-edge tech.
Liu’s innovations bridge cultural gaps, enhance accessibility, and equip educators for the diverse classrooms of tomorrow.
Revolutionizing Science Education with Immersive Technology
Immersive technologies like AR and 3D models are transforming science education by making learning more accessible and engaging. These tools provide hands-on experiences without the need for costly labs, allowing students to visualize complex concepts and gain deeper understanding. Virtual labs and biomedical illustrations help overcome financial and logistical barriers, enabling more students to explore STEM fields. AI-driven platforms also personalize learning, boosting confidence and interest in science. Best practices for using these tools include setting clear objectives, offering real-world applications, providing concise modules with feedback, and utilizing multi-modal learning strategies. By integrating these innovations, educators can bridge gaps in science education, enhance engagement, and inspire the next generation of scientists.
AI News
The AI Revolution Sparks Urgent Innovation in the Power Sector
As AI’s energy demand skyrockets, utilities are being forced to accelerate innovation like never before. Data centers could consume over 9% of U.S. electricity generation by 2030, driving the need for cleaner, more efficient energy solutions. Traditionally slow-moving, utilities now face pressure from Big Tech’s aggressive net-zero goals, pushing them to scale renewable energy, optimize power grids, and embrace AI-driven efficiencies. AI itself is emerging as a critical tool in power management—optimizing solar and battery usage, streamlining permitting, and even preventing power outages. This shift isn’t just about keeping up with AI’s energy demands—it’s about transforming the entire utility industry into a smarter, cleaner, and more resilient force for the future.
Microsoft Expands AI Hub to Boost NPU-Powered Applications
Microsoft has launched an upgraded AI Hub in the Microsoft Store, featuring applications optimized for Neural Processing Units (NPUs) in CoPilot+ PCs. These specialized AI chips, designed to handle AI-intensive tasks like image recognition, content creation, and voice transcription, aim to enhance efficiency and system performance by offloading workloads from the CPU and GPU.
Despite the initial lack of NPU-optimized apps, Microsoft is now actively encouraging developers to build software that leverages AI PCs. Apps like Camo Studio, which accelerates webcam visual effects, are already utilizing NPUs on select laptops.
Industry experts note that while AI PCs are becoming more appealing, Microsoft will need a larger catalog of apps to drive adoption. Developers are now receiving support from Qualcomm and Intel, both of which are offering tools and resources to encourage NPU app development.
With features like Recall, Click to Do, and semantic search on the CoPilot+ roadmap, Microsoft and its partners are positioning AI PCs as the future of computing, bridging the gap between hardware and software for next-gen AI applications.
Humane Bricks AI Pin as HP Acquires Company for $116 Million
Humane’s AI Pin, once hailed as the future of AI wearables, is officially dead. HP has acquired the struggling startup for $116 million, a fraction of its previous $850 million valuation.
Key takeaways:
Humane AI Pin is discontinued – Devices will stop functioning on Feb. 28, and owners have until Feb. 27 to claim a refund.
HP acquires Humane’s CosmOS AI operating system, patents, and talent – Plans to integrate CosmOS into PCs, smart printers, and conference room devices rather than wearables.
Humane’s founders, Imran Chaudhri and Bethany Bongiorno, will lead a new AI division at HP, focused on AI-driven computing rather than hardware.
Why did Humane fail? Reviews criticized the AI Pin’s slow responses, inaccuracies, and limited usefulness, along with a fire hazard recall for its charging case.
HP’s acquisition signals a shift away from AI wearables and toward AI-driven enterprise and productivity solutions. Meanwhile, AI hardware remains a work in progress, with Meta and other companies still pursuing AI wearables in 2025.
Netflix Faces Backlash for Using AI to Clone Gabby Petito’s Voice in Documentary
Netflix’s latest true-crime documentary, American Murder: Gabby Petito, has sparked controversy after filmmakers used AI to recreate the voice of the murdered travel vlogger to narrate her own journal entries.
Key Takeaways:
AI-generated voice of Gabby Petito – The documentary team used "voice recreation technology" to have Petito seemingly narrate her own writings.
Public outcry – Many viewers found the decision “deeply uncomfortable,” “disrespectful,” and “monstrous.”
Netflix claims it had family approval – The Petito family reportedly gave their blessing, and the filmmakers defended the move as a way to “tell the story through Gabby’s words.”
Ongoing Hollywood AI controversy – This follows recent backlash over AI-modified voices in The Brutalist, prompting the Academy to consider AI disclosure rules for films.
Using AI for historical narrations isn’t new, but recreating a recent murder victim’s voice raises ethical concerns about AI’s role in entertainment. Should Hollywood disclose AI usage in documentaries, or should some voices remain untouched?
AI Tools
Rabbit: AI-Powered Reading Assistant for Faster Learning
Rabbit Reader is an AI-powered reading assistant designed to simplify complex texts, making it particularly useful for students, researchers, and professionals. It utilizes advanced AI models, including GPT-4o and Sonar Pro, to provide contextual explanations and sources for deeper understanding. Originally created to enhance reading comprehension of Arxiv papers, it now helps users decode difficult content across various online texts. Rabbit operates on a subscription-based model with no free trial but offers a 7-day money-back guarantee. Users can cancel anytime while maintaining access until the end of their billing period.
Open eLMS AI: Transforming eLearning with AI-Powered Course Creation
Open eLMS AI is an advanced AI-driven eLearning platform enabling users to instantly create professional-grade courses. Designed for businesses, educators, and training providers, it automates the entire course development process, generating scripts, animations, images, voiceovers, and video presenters in seconds. Users can edit courses with an intuitive drag-and-drop interface and export them as SCORM-compliant files for easy integration with any LMS. With over 21 years of expertise in eLearning, Open eLMS AI ensures high-quality, accessible, and scalable content creation. A 14-day free trial allows users to explore its features risk-free.