Generative AI

Qualcomm China Chairman Jerry highlighted at the 2025 Gmosemiconsductor Conference how terminal-side AI is revolutionizing the semiconductor industry. He discussed AI becoming the new user interface, driving structural demand for chips, and the critical hardware-software synergy required for on-device AI performance. Jerry also emphasized Qualcomm’s collaboration with China’s ecosystem partners to accelerate AI adoption across smartphones, PCs, and vehicles, recognizing China’s leading role in AI innovation.

MingLue Science and Technology launched AdEff, an AI-powered platform for global advertising testing and optimization, at ESOMAR “Renewal Tour 2025” in Singapore. Built on the HyperGraph Multi-Modal Large Language Model (HMLLM), AdEff quickly predicts consumer reactions to ads, offers improvement recommendations, and aims to revolutionize marketing productivity. The platform integrates proprietary data and a mixed-expert model to provide faster, cost-effective, and more relevant content measurements, with trial use globally available starting June 19th. Continuous enhancements, including expanded content types and global expansion, are planned.

At MWC Shanghai 2025, Huawei’s rotating chairman emphasized 5G-Advanced (5G-A) and mobile AI’s transformative potential. Rapid 5G-A deployment and AI device adoption are accelerating. Generative AI is experiencing explosive growth, reshaping user experiences. The convergence of AI and IoT is generating significant industrial opportunities, with connected vehicles and robots poised for massive expansion by 2030, predicting a tenfold increase in AI-IoT connection volumes and value.

Meta is in advanced talks to invest over $10 billion in AI data labeling startup Scale AI. This deal, if finalized, would be a historic private financing, reflecting the importance of data in the AI landscape. Scale AI, which provides data labeling services crucial for training machine learning models, has experienced rapid growth. Meta’s investment signals a shift toward external AI partnerships and supports its strategic focus on AI, including its Llama language model and AI-powered integrations across its platforms.

AI-generated deepfakes are enabling sophisticated scams by impersonating public figures and contacts, with global fraud losses exceeding $12 billion in 2023. Using facial recognition, biometric extraction, and neural networks, modern tools create real-time fake videos with over 90% accuracy, narrowing detection opportunities. Experts note subtle flaws—unnatural blinking, lighting mismatches, or blurring—as red flags. Stanford’s Dr. Elena Voss advises challenging unexpected video requests with sudden movements to expose rendering limitations. Corporations are deploying AI to analyze micro-expressions and blood-flow patterns, while individuals are urged to verify requests via secondary channels and adopt biometric authentication. The fight against AI-driven deception hinges on “intelligent skepticism” and advanced detection technologies.

The ISG Provider Lens™ report (September 2025) analyzes how AI, including GenAI and agentic technologies, transforms finance and accounting outsourcing (FAO) by shifting vendors from back-office roles to strategic partners. Across 48 providers in four domains—invoice-to-pay, order-to-cash, R2R/tax, and FP&A—AI automates workflows, reduces manual tasks by 60%, enables real-time insights, and supports predictive financial planning. The study evaluates innovation, global adoption, and vendor capabilities, emphasizing AI co-pilots, blockchain integrations, and tools turning spreadsheets into dashboards, to help enterprises benchmark partners reshaping CFO priorities and operational agility.