Artificial intelligence-human AI collaboration
The COVID-19 outbreak has created uncertainties and changes in all aspects of daily life and work. It also pushes business leaders and factory owners to act faster than ever, to survive and outperform others. This provided more possibilities for AI technologies and accelerated the deployment of AI in manufacturing.
From a practical perspective, we will see more AI-based applications specifically targeted at various aspects of decision-making, such as improving production/operational efficiency, reducing downtime, providing predictive maintenance, optimizing the supply chain and reducing energy consumption.
Human and AI capabilities complement each other. Humans are creative, can see above the task at hand, and can apply knowledge from other experiences in a task at hand. However, humans are less perfect at repetitive tasks. AI is efficient and diligent, but less creative. There is great potential for collaboration between humans and AI, and humans need to interact collaboratively with AI as they continue to fulfill the roles of project owners, system trainers, end users and will interact with AI throughout the lifecycle of a project.
AI is applied to a wide range of applications and industries. In process control, AI and machine learning are applied to advanced process control (APC) applications and autonomous factory operations. In discrete manufacturing, AI is applied to robotics. All processes could theoretically be controlled by some form of long-term artificial intelligence. Regardless of the application, there are a few basic steps identified by ARC to ensure the success of your AI deployment:
Based on presentations from the 2022 ARC European Forum, this strategic report provides the latest AI use cases and a comprehensive comparison of the strengths and challenges of humans versus AI. This report outlines how humans and AI can collaborate effectively and what key dimensions and applications need to be considered. Companies that have supported the event include Microsoft Project Bonsai, Dow Chemical, NNaisense, ABB and Throughput AI.
All use cases
The following industry use cases for AI technologies were all shared at the European ARC Forum 2022.
Microsoft Project Bonsai shared their views on autonomous systems and how the Bonsai platform can optimize equipment and processes by sensing and reacting in real time. According to Microsoft, Autonomous Transformation is an evolutionary process consisting of four stages:
- Human intelligence as control only: systems have no intelligence of their own).
- Intelligent Insights: Human operators are likely guided by data-driven insights.
- Support capability: Systems have some analytical capabilities.
- Autonomous operations capability
From the second stage, companies can use supervised and unsupervised learning to achieve things like better predictive maintenance and demand forecasting. The Bonsai platform combines simulation, application deep learning and machine learning. Bonsai can help users create AI models with their own experience and industrial know-how, and accelerate the development of the last three stages. Here are some Bonsai case studies:
- PepsiCo uses a reinforcement learning model trained in a snack production line for decision support. The model equates to a moderately expert operator to handle changing conditions of the plant environment and incoming resources and helps more junior engineers accelerate their learning journey. Deeper application learning was also used in HVAC optimization, achieving energy savings of 5-15%.
- Bell Flight used a trained drone control officer to perform vision-based precision landings on verified targets in a GPS-less environment.
Dow Chemicals shared how the cloud enabled predictive maintenance through the optimization of manufacturing data and analytics. Unscheduled downtime represents huge lost revenue for process industries. Leveraging predictive maintenance to act before a breakdown occurs is one of the main applications of AI in manufacturing.
In Dow Chemical’s own project, three separate people work as a team. Data scientists train the models and verify the Azure Cloud environment. Azure Cloud developers ship data from the factory to the cloud and feed the models with data. On-site operators perform monitoring, take action when they receive information from the systems, and provide feedback into the systems. This cycle of work continues continuously to build algorithms and models more efficiently. In this project, cloud computing is needed to solve a few headaches, including:
- Access to data: Collecting and integrating separate data/information between organizations is a big challenge.
- Training and deploying iterative models: All activities, including coding, testing, releases, deployment, and monitoring, are done sustainably.
- Flexible and scalable: Tested AI models can be deployed to other equipment or processes with readily available sensors.
Debit AI shared how AI helps optimize the supply chain and convert visibility into actionability. The supply chain is opaque, piecemeal and inefficient all the time. The COVID-19 outbreak has made its bottlenecks more evident than ever in recent years. Meanwhile, companies are facing increasing pressure to improve supply chain efficiency and find opportunities for better business, operational, financial and sustainability outcomes and outperform industries over the long term. As an enabler, AI can help leverage existing data and the domain expertise of existing teams, to increase production, inventory turns and profitability, minimize excess inventory and reduce waste and waste. CO2. Some cases shared here:
- Improve SKU (Stock Keeping Unit) flows: achieved annual systemic optimization revenue and increased savings of over $20 million, optimized SKUs of over 250,000 real-time units with an OTIF (On Time and In Full) rate of up to >95 % (vs.
- AI multi-dimensional segmentation: Intra-tier client movements can be identified automatically, with the best and worst performers flagged.
- Improve logistics: AI software, including demand detection AI, segmentation AI, and prediction AI, can help determine high-demand and optimally routed products, automatically analyze and recommend promotions to eliminate inventory. A client in the building materials and construction industry realized an additional profit of $31 million, a net financial impact of 3.3% and less waste of non-recyclable materials and CO2 emissions (~ 226,000 tons and 104,000 tons, respectively).
ABB shared details about its AI-assisted operator to illustrate how AI could help operators. Production processes generate many alarms, and it is easy for operators to fall victim to alarm flooding or to ignore alarms that may be critical. AI is one of the latest technologies that can provide operators with more direct, efficient and cost-effective support for effective long-term alarm management. AI-assisted operator is a mix of different AI applications, its workflow includes:
- Anomaly detection: Based on the analysis of historical data and previous issues, the AI platform will classify them and predict what is happening and what is wrong.
- Root Cause Analysis: Several dozen tags can be traced downstream in the topology-based method, to find the real alarm trigger.
- Recommended actions: The recommendation engine will indicate exactly what to do or what effective actions have been done previously in similar scenarios.
- What-if model: A quick simulation can be run in seconds to see what will happen before taking action.
- Knowledge Extractor: Years of operator actions and events could be verified, which may look like current cases, with only a few standard deviations.
Nnaisense presented details of its AI-based adaptive control system called Adaptive, Rational Core. The control system will build a modular, task-independent and verifiable model. Different from traditional reinforcement learning, the system can be imposed with new goals and constraints at any time and perform tasks on command. The system can not only control the plant, but also predict the future in various scenarios and plan backwards in time from the target state to the current state.
Human and AI collaboration happens all the time. All AI-based systems must go through many iterations of validation and verification alongside continuous interaction with factories and the outside, or when taking on new tasks or missions. Humans will go through five stages when interacting with AI-based systems:
AI adoption phase
Most people may be curious about how AI technology is progressing in practice. According to the ARC Forum survey, ‘early products and solutions available’, ‘R&D’ and ‘beta testing/selected projects’ are the common milestones among participants. Generally, it is still in a nascent stage.
As our expert roundtable revealed, companies early in their AI journey struggle to define a strategy for a successful technology implementation. Many companies in the technology assessment phase opt for a trial and error approach. However, a strategy is essential to formulate KPIs and may be necessary to secure funding from management and continue the AI journey. As these journeys can go on for quite a long time, it is important to have the right strategy, goals and KPIs as management teams can change which can lead to the threat of shutdown.
- Executive Overview
- All use cases
- AI adoption phase
- Humans versus AI: strengths and challenges
- AI process control
ARC Advisory Group customers can view the full report on the ARC Customer Portal
If you would like to purchase this report or obtain information on how to become a client, please contact us
Comments are closed.