In a recent interview with CIOTechOutlook, Charles Daniel Gnanesh, Head of IT, KONE, shares his views on how Generative AI is used to enhance quality control, seamless integration of AI technologies in minimizing disruptions, potential risks involved with implementing generative AI and more. Charles has designed plethora of software both on the application and system side using cutting edge tools and managed diversed set of resources.
How crucial is it to establish a well-defined strategy and grasp the implications before effectively deploying Generative AI?
It is essential to have strategic thinking in the organization. Ideally, the board should drive the initiative and ponder how generative AI can be utilized in the firm so that the workforce gets a brief understanding of working with this new technology. Secondly, there must be benefit mapping for every innovation. For instance, employees can suggest ways to adopt technologies, new working styles, tools, etc. However, it's all about how customers benefit, which is more important. The other is the road map; it involves what activities are to be performed the timeline, the required investment, the sources of funding, and the expected return on investment. It's a valid point. Besides, there is a need for strong governance. To exemplify, the AI chatbot can be adopted without complications, but we do not know the consequences of its being in public domain. Hence, strong governance is pivotal in every implementation that is done in the company.
Finally, there needs to be a large AI community within the firm that talks about what is right, where this technology trend is heading and new laws. Hence, to adopt this technology, these are some of the points that need to be focused on as part of the well-defined strategy.
Ensuring high quality in manufacturing is crucial, what are the ways in which Generative AI can be used to enhance quality control in manufacturing?
The manufacturing industry has emerged into the digital revolution, with many automated processes. However, many manual interventions are required, mainly in the elevator industry. Similarly, a manual assembly line is going on, and documentation is created manually for invoicing, tendering, securing an order and AMC. Hence, many activities are still performed manually, and errors exist. While adopting new technology, organizations need to ensure that errors are minimized so that whenever a manual intervention is required, the machine will notify the user in case of any wrong input, considering the existing information provided. Hence, this is one of the ways Generative AI can help them. The AI is recursive by nature, so it can reinforce the model's effectiveness day by day by making the required changes. As the model evolves into a better engine, errors can be reduced, contributing to fewer manual errors.
How important is the seamless integration of AI technologies in minimizing disruptions during adoption?
There is a lot of massive data involved concerning generative AI. It utilizes a generative adversarial network (GAN), which requires numerous data processing processes so that it can be possible to ascertain what to respond and whether it is correct or not, based on probability. Seamless integration is critical to identifying and using the correct data in the organization. There is no need to reinvent the existing data, and it can be a large data lake that needs to be looked into, pick up the right transactional or non transactional data and build again. Hence, once the GAN is ready, it becomes autopilot, which will keep reinventing itself to get a good model. Besides, it is essential to integrate with the data lake or data layer in the firm so that AI becomes more accessible to work with.
Evaluating risk management capabilities is essential. What are the potential risks associated with implementing generative AI in manufacturing, and how can these risks be mitigated?
The generative AI risk applies to any industry; when it comes to manufacturing, there are operational technology environments in the factories. Large machines that automatically paint or dry are highly likely to be prone to attacks and connected to the Internet. So, cybersecurity is one of the significant risks while implementing something outside of factories. For instance, for AI support, there is a need for a large model, which might be on the cloud, and these operational technology instruments or devices might need to connect to it. And then, it opens up something where we need to pay attention to ensure that it doesn't affect the factory operations. The other risk is that when you train a model, it could be trained as bias model. For instance, when the model is trained in India, a similar model needs to be deployed in Europe or America; it might need to work differently. However, since it is already trained for India, this bias can creep in concerning Generative AI, which is always a risk.
Eventually, with insufficient data, you will end up making wrong decisions using AI, which is always a potential risk. As the model matures, it will start giving the appropriate decision-making information, and it's always good to wait for the good model to arrive prior to production and proceeding with it.
What emerging trends in Generative AI do you believe will have the most significant impact on manufacturing in the next 5 years?
Governments and technology personnel are increasingly recognizing the importance of generative AI, and stringent laws and regulations are expected to be launched across several industries, including manufacturing. The laws aim for collaboration, ensuring all parties work towards mutual benefits while following ethical standards. The next trend is that there are a lot of emerging micro-factories. For instance, in the manufacturing industry, assume there is an assembly line in which a car can be assembled or an elevator from one point to the other and then moves to various stages, ending up with an elevator or a car. However, there is a trend towards setting up micro-factories in this modern era. A micro-factory can be set up close to the location where it will be used. It occupies minimal space, enabling cars to be built at a particular spot. Required materials are delivered to this spot using robots and
AI technology. The vehicle is assembled there before moving to the next place for the next assembly, and this approach eliminates the need for large assembly line-based factories in the future.
The next trend is the audio user interface. For
Gen AI technology, prompting is critical. Wearable computers with human language interfaces will soon be available, where users can just keep walking and conversing with the AI assistant on the computer. The computer can also be connected to the internet, and users can look for information. Hence, the audio user interface is an effective trend emerging in the manufacturing space.
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
