Asia Times: Do Japanese manufacturers use ChatGPT?
ChatGPT: It is possible that some Japanese manufacturers use ChatGPT or other similar language models for various applications… However, I do not have access to specific information on which Japanese manufacturers are currently using ChatGPT or other similar language models…
Asia Times: How about Panasonic?
ChatGPT: I do not have access to information on whether Panasonic specifically uses ChatGPT or other similar language models. However, it is possible…
Asia Times: How about this?
Nikkei Asia: Panasonic unit deploys ChatGPT-style AI to improve productivity (March 8, 2023) OSAKA/PALO ALTO, U.S. – A system development company under Panasonic Holdings has begun large-scale deployment of easy-to-use, interactive artificial intelligence that it hopes will improve productivity.
Panasonic Connect has made it possible for all its 12,500 employees in Japan to use the AI as an “assistant” for tasks such as drafting documents, Nikkei has learned.
Through a partnership with Microsoft, the company has developed an AI assistant that answers employees’ questions, utilizing technology from OpenAI, a U.S. company that developed ChatGPT, an interactive AI….
The company uses the AI system to produce statistical data in a format that can be easily handled by spreadsheet software, to create templates for documents, and to ask for ideas on how to proceed with meetings….
The system can be used only within Panasonic Connect’s internal network to prevent information leakage. To reduce such a risk, Panasonic Connect has agreed with Microsoft not to use any content for secondary purposes such as AI learning. In addition, the company has warned employees not to input personal or confidential information.
ChatGPT: Thank you for the update! It seems that according to the article on Nikkei Asia, Panasonic Connect, a system development company under Panasonic Holdings, has indeed deployed an AI assistant that uses technology from OpenAI, the developer of ChatGPT, to improve productivity….
Beware
Beware, short-cut seekers. The preceding conversation shows how the verbose ChatGPT, even when prompted with the name Panasonic, failed to notice an article published five weeks ago on a well-known English-language business website. Google found the article in less than a second.
And from the article we learn that – before Elon Musk, Steve Wozniak and more than a thousand other concerned persons issued a March 14 “call” on all artificial intelligence labs to “immediately pause for at least six months the training of AI systems more powerful than GPT-4” due to “profound risks to society and humanity”– Panasonic had already started to implement a low-risk internal version of the technology.
So we’re caught up on the furor over ChatGPT, the most widely publicized information on the subject of AI in Japan – but in the process that furor has distracted us from more significant applications across a range of industries. Examples include:
- Pharmaceuticals: Astellas Pharma has created a cellular drug discovery platform integrating humans, AI and robotics that it claims can shorten the time of drug candidate compounds acquisition by approximately 70% in successful cases. It is now being used in a joint research project with the Center for iPS Cell Research and Application of Kyoto University to “promote the utilization of differentiated cells and tissues derived from human induced pluripotent stem cells (iPS cells) and to create innovative medical solutions.”
- Regenerative medicine: Canon plans to apply its deep-learning neural network for image processing to regenerative medicine following its acquisition of the high-speed cell cultivation system technology of Kyoto Seisakusho (scheduled to be completed in May).
- Factory worker support: Hitachi has developed a “real-time work deviation detection” system based on deep learning that reduces inefficiencies resulting from the interaction of humans and machines in highly automated factories. The company is also using neural networks to improve fault detection in complex industrial and infrastructure equipment, systems and networks. Once primarily an industrial hardware conglomerate, Hitachi is now applying digital technology to everything it does.
- Supply chain logistics: Hitachi has also developed algorithms to maximizing the efficiency of complex supply chain networks involving multiple factories, warehouses and shops linked by land, sea and air. This includes production scheduling, inventory management and delivery route efficiency maximization.
- Quality Control: NEC uses machine learning to standardize and speed up product inspections, saving time once spent on visual inspection and improving accuracy through objective data analysis. NEC also uses automated real-time analysis of temperature, vibration, pressure, voltage, flow rate and other data from sensors to monitor power generation facilities. This “enables anomalies to be detected at an early symptom stage, including subtle anomalies which would not be noticed even by inspection experts.”
- Industrial Cyber Security: Fujitsu provides services ranging from the assessment of a manufacturer’s operational technology to cyber security system design and implementation, continuous monitoring, incident investigation and system upgrades. Based on data analytics, they derive from Fujitsu’s own experience as a producer of computing and communications equipment.
- Embedded systems: Semiconductor maker Renesas has put artificial intelligence on embedded systems (e-AI) in order to run inference on its MCUs, MPUs and AI accelerators. This enables real-time processing without the network delay inherent in cloud computing. Applications include factory automation monitoring (to deal with, for example, abnormal vibration), automated system device coordination and motor control or other failure prediction.
- Semiconductor metrology: Hitachi is applying deep learning to critical dimension scanning electron microscopy (CD-SEM) in order to more precisely measure defects in high-aspect-ratio holes in 3D semiconductor devices (e.g., NAND flash memory) after etching and thereby improve process control. Hitachi also makes etch equipment.
The list goes on and on, covering healthcare, traffic control, waste management, wind turbine performance – anything that requires the processing of enormous amounts of data. All of this is way past ChatGPT, although that doesn’t mean that it might not take your job.
Definitions
IBM points out that “Perhaps the easiest way to think about” artificial intelligence, machine learning, neural networks – which mimic the human brain through the use of algorithms – and deep learning “is to think of them like Russian nesting dolls. Each is essentially a component of the prior term.”
In other words, “machine learning is a subfield of artificial intelligence. Deep learning is a subfield of machine learning and neural networks make up the backbone of deep learning algorithms.”
Arm, the IC design company, tells us that:
AI Inference is achieved through an “inference engine” that applies logical rules to the knowledge base to evaluate and analyze new information. In the process of machine learning, there are two phases.
First, is the training phase where intelligence is developed by recording, storing, and labeling information. If, for example, you’re training a machine to identify cars, the machine-learning algorithm is fed with many images of different cars the machine can later refer to.
Second, is the inference phase where the machine uses the intelligence gathered and stored in phase one to understand new data. In this phase, the machine can use inference to identify and categorize new images as “cars,” despite having never seen them before. In more complex scenarios, this inference learning can be used to augment human decision making.
By the same token, if you feed a news article to ChatGPT, it will quote it right back at you a moment later. But it won’t be running the industrial economy.
Follow this writer on Twitter: @ScottFo83517667