InnoPower.ai Daily AI Newsletter – 10/16/2025

MIT assistant professor and J-WAFS researcher Ali Aouad is advocating the use of an algorithm to revolutionize food assistance policy in the Global South. This approach aims to leverage digital platforms for optimizing food subsidies, thereby enhancing nutrition and supporting sustainable agricultural practices.

Jack & Jill, leveraging conversational AI technology, intends to streamline the job search process by eliminating the traditional reliance on listings and resumes. With a $20 million seed funding round secured, the company is poised for a strategic expansion across the United States. This development signifies a notable innovation in the recruitment sector, utilizing advanced AI to facilitate job hunting.

The co-founders of Eightfold have launched Viven, an AI startup focused on creating digital twins to simulate interactions with unavailable co-workers, securing $35M in seed funding. The investment round was led by Khosla Ventures and Foundation Capital, highlighting the growing interest in leveraging AI to enhance workplace efficiency and communication.

Liberate, a company that integrates AI into the insurance sector, has successfully raised $50 million at a $300M valuation to enhance the efficiency of back-office operations in Property & Casualty (P&C) insurance through AI automation. This funding will enable the expansion of its AI agents across various sectors including sales, service, and claims, aiming to streamline workflows and reduce manual tasks within the insurance industry.

Meta is collaborating with semiconductor company Arm to improve its artificial intelligence capabilities, indicating a significant upgrade to their AI systems. This partnership is part of Meta’s broader initiative to expand its infrastructure, underscoring the social media giant’s commitment to advancing AI technology within its platforms.

Anthropic has launched Claude Haiku 4.5, an AI model offering significant advancements in speed and cost-efficiency over its predecessors, at one-third the price and double the speed. For the first time, Anthropic is offering free access to this advanced AI on its Claude.ai platform, a move aimed at democratizing AI access and intensifying competition with major players like OpenAI. This strategic shift could significantly impact the enterprise AI market.

Visa unveils its Trusted Agent Protocol to distinguish legitimate AI shopping assistants from malicious bots amidst a 4,700% surge in AI-driven traffic to U.S. retail websites. This protocol allows merchants to verify the authenticity and intent of AI agents, aiming to secure the burgeoning field of agentic commerce and protect against sophisticated bot-induced fraud.

A new method has been developed to enhance the ability of generative AI models, specifically vision-language models, to accurately identify and locate personalized objects in unfamiliar settings. This training technique improves the models’ efficiency in recognizing distinctive items within diverse scenes, making them more adept in navigating and interpreting new environments.

The article commemorates Professor Emerita Jeanne Shapiro Bamberger, who passed away and was celebrated for her pioneering work in integrating artificial intelligence into music education. As a former department chair, she played a crucial role in innovating how children learn music through the use of AI, marking significant contributions to both music education and AI application in learning.

Kimaya Lecamwasam, a PhD student at the Media Lab, is at the forefront of merging neuroscience, AI, and music to engineer novel mental health solutions. Her research focuses on leveraging the therapeutic potential of music to influence mental well-being positively, highlighting a multidisciplinary approach that converges different fields to address mental health challenges.

EAGLET, a new academic framework, has been developed to enhance the performance of AI agents on long-horizon tasks by offering a ‘global planner’ for better task planning. This innovation helps reduce errors and increase efficiency without requiring additional data labeling or retraining, addressing the challenge AI agents face with tasks that span multiple steps and extended periods.

Anthropic has launched Claude Haiku 4.5, the latest iteration of its compact AI model, promising to deliver performance comparable to its larger Sonnet 4 model at a significantly reduced cost and increased efficiency. This development represents a stride towards more accessible and efficient AI technologies, making sophisticated AI tools more attainable for a wider range of users and applications.

A new AI tool named SpectroGen acts as a ‘virtual spectrometer,’ offering the capability to generate spectroscopic data across various modalities such as X-ray or infrared. This innovation enables rapid assessment of material quality, streamlining the process significantly.

Google has launched Veo 3.1, an advanced AI video generation model, enhancing capabilities in narrative control, audio integration, and video realism. This release not only broadens opportunities for hobbyists and content creators but also presents significant advantages for enterprises and developers with its higher quality, customizable features, and integrated audio options for a seamless video creation experience.

The Dfinity Foundation has launched Caffeine, an AI platform that enables the creation and deployment of web applications using only natural language prompts, eliminating the need for traditional coding. This platform operates on a decentralized infrastructure tailored for autonomous AI development and aims to provide a comprehensive alternative to human technical teams by harnessing AI to manage all aspects of application development and maintenance.