Advantech and Fujitsu bring AI innovations to the workplace

On December 15, 2024By Littum

The use of artificial intelligence (AI) is gaining tremendous momentum across various sectors, especially in workplace technology. With companies recognizing the potential to increase efficiency and transform routine operations, innovations are rapidly shaping the future of how we work. Among the latest breakthroughs, Advantech recently introduced two new AI acceleration modules – EAI-1200 and EAI-3300 – designed to significantly improve AI performance without sacrificing energy efficiency. Together with Hailo’s advanced Hailo-8 processors, these modules promise high computing power of up to 52 trillion operations per second (TOPS) and boast more than twelve times the energy efficiency compared to conventional AI cards and GPU cards.

These acceleration modules are available in standard M.2 and PCIe formats and can be easily integrated with various x86 and ARM-based platforms. This means existing systems can be quickly upgraded to incorporate AI capabilities, allowing developers to run inference tasks more efficiently. The modules also have easy-to-use software tools like the Edge AI SDK, which seamlessly improves the integration process with applications powered by HailoRT. For developers, this option offers the ability to optimize resource allocation efficiently by performing application processing primarily on the CPU, which could be a game-changer for industries that rely on vision AI.

On the other side of the globe, Fujitsu has been making waves with its ground-breaking video analytics AI agent that aims to boost security and operational efficiency in frontline workplaces. This innovative AI agent uses data from workplace cameras along with written documentation to generate reports and suggest improvements based on live video footage analysis. Designed to act as the backbone of Fujitsu’s AI service, Kozuchi, the new technology boasts incredibly accurate video processing capabilities, even with longer video durations. The company plans to begin testing this AI agent in fiscal year 2024 and expects to roll out full implementation by January 2025.

The AI works through a sophisticated large language model (LLM) that learns to recognize the 3D spaces captured on video. It integrates this spatial analysis with written information such as safety guidelines, ensuring that AI retains relevant data to significantly improve analysis quality. The remarkable accuracy of AI’s assessment and its ability to understand long-term recordings places Fujitsu at the forefront of workplace technology.

A key aspect of AI is its ability to retain contextual memory from video content, which improves video understanding by instructing the model to focus on specific behaviors, such as identifying unsafe behaviors among employees. This selective analysis model enables companies to maintain effective safety protocols and monitors behavior classifications, which are increasingly necessary as workplace safety becomes more pronounced.

Fujitsu has partnered with Carnegie Mellon University to monitor AI’s development within its new FieldWorkArena – an evaluation platform specifically created to test and improve agent capabilities. FieldWorkArena offers real-world scenarios, complete with video feeds and documentation, allowing for extensive evaluation to refine agent computing skills. Researchers will have access to this evaluation environment, and assignments will be made available on GitHub and Fujitsu’s research portal.

Both Advantech and Fujitsu exemplify the rapid evolution of workplace technologies and their unwavering commitment to redefine productivity through AI. Innovations such as the EAI-1200, EAI-3300 and Fujitsu’s Video Analytics Agent mark significant progress towards integrating AI across various sectors, improving not only operational efficiency but also employee safety and performance. What remains to be seen is how these advances will affect employer choices and how workplaces will adapt to accommodate such innovative technologies.

Looking to the future, the continuous integration of artificial intelligence into workplace systems indicates promising prospects for industries trying to optimize workflows, strengthen safety measures and streamline operations. The advances reflect broader trends where companies increasingly value technological agility and aim to remain competitive by deploying tools capable of tackling modern workplace challenges.