Artificial Intelligence: Where Shall We Go Tomorrow?
Published on : Sunday 03-03-2024
In the concluding part of the 3-part series, PV Sivaram examines areas where AI is stepping in manufacturing.
The question indeed applies to humans and technology – where shall we go tomorrow? Actually the question becomes poignant because Human Intelligence and Artificial Intelligence have to coexist. It is human intelligence which develops AI. But human intelligence also views AI as a competitor! Since it is an ever-striving goal of manufacturing to improve in terms of the various indices like throughput, efficiency, quality, safety, etc., it does look to be a natural relationship between AI and the industry.
All future gazing is rather hazardous, given the speed of development of technology, and the interaction of different technologies with each other. So we will restrict ourselves to near future, and only put out some dreams of a golden far future.
State of relationship today
The word ‘Artificial’ brings up a range of negative connotations – unreal, undependable, cheap, insincere, fake, illusory, mechanical, arbitrary, etc. Artificial intelligence somehow inherits these feelings and brings up a defensive reaction. But manufacturing itself is largely a process to produce artificial items; the linguistic connotations are not so much a road-block.
As it is today, AI is not the heart and soul in manufacturing. As yet manufacturing has not formed a clear picture of the relationship it wants to forge with this new technology. There is much hype about potential and possibilities, but poor articulation of application use cases. There is work to be done, both by AI and by the industry.
Visual image of AI
People and organisations have projected AI to be many different things. Somewhere it is described as a co-pilot, thereby invoking an entity well equipped with all knowledge and capability of a pilot, however just waiting at the elbow to pitch in when called for, and otherwise staying out of the way.
There is another view of AI as a Jeeves to a bumbling Bertie Wooster, who miraculously appears as and when the master has a need for help, even if he himself does not know it or ask for it.Yet another view is of a ghost out of the bottle, who can answer any question and perform any task.
I would liken it to the role of electricity, which is twofold. It enhances speed and performance of every tool, when switched on. It also gets embedded in many tools to enhance performance unobtrusively and makes work more easy and accurate like a power-steering or an automatic gear shift. With this view, AI is actually a good fit for helping out manufacturing processes.
What makes AI gain power day by day
Progress in AI started in the early eighties after the so-called AI-winter. It has reached a particular momentum in the new millennium due to a variety of factors, seemingly unrelated to each other. But definitely these factors have a mutual correlation. Some of the factors are – proliferation of computing devices in both industrial and personal space, connected devices – internet, IoT, always online, proliferation of data and harvesting of data.
Algorithms and data and chips
AI is commonly imagined to be a set of algorithms – a mass of computer SW. This view is not wrong. AI is given rules by human programmers, to respond in specified ways to data which is provided to the AI device. While this is true and correct, it is not the full picture. AI also has the capability to learn from its experience, by working with many different sets of data in the ‘real’ world (mostly it is a cyber image of the real world), to derive different decision paths. In that sense AI is more than a set of IF-THEN-ELSE constructs. The third component is the hardware. At a full functional capacity, AI needs very high computational power. Development of chips engineered precisely to cater to AI becomes a game changer. By placing enormous computational power, the sort of applications which AI can be used to solve rapidly increases.
The combination of these three factors increases the ‘power’ of AI day by day – we mean the application areas where there would be no alternative to AI.
Race for domination
Power of nations or groups or companies is largely from the economic might. Economic might is obtained by control of access to natural resources, efficient production of items which will be in demand, and accurate prediction of future demand. All of this can be imaginably better, and thereby the groups become mightier, by deploying AI. Therefore AI itself becomes a resource for which there is going to be competition.
If we look at the AI tools and processes as a crop, and the developer activity as cultivation a particular picture emerges. As of today, wealthy nations with foresight actively fund and encourage development activity. As well skilled technical talent is an important ingredient, they encourage better quality workers to produce AI for their purposes. The other material input needed is the HW, viz., chips, which go today by the name of GPUs. Chip development and production is also highly capital intensive and therefore available in few elite lands.
Third aspect, which is very important, is data. AI is a data guzzler. AI needs much data during development of algorithms. When the algorithm or model is developed, data is again needed for verifying or validating the algorithm. Data is actually generated by consumers of a product or service. There is a scramble to get at this data by countries and corporations. Large amount of data is obviously created in places with high populations. So access to data becomes an important goal in global diplomacy.
For example in developing AI for medical purposes, much patient data is needed regarding consumption of a particular drug and the resultant effects and side effects. This data for some kinds of treatments are most prolific in teeming millions of Africa and Asia. Giant corporations vie with each other to secure this data.
So in this race, the future of AI is tilted towards the haves and away from have-nots. The have-nots may have to buy the AI tools and products which their own people and data have worked to develop. Else they may have to live and work with more imperfect tools and suffer a competitive disadvantage.
Competent servant or jealous master?
How would industry and AI live with each other? Industry will treat AI as a set of tools, similar but more powerful than tools of the previous generation. It will need some imagination to reengineer the process when such powerful tools are available freely.
How AI will help manufacturing industry
We could frame the question differently. Where does the manufacturing industry need AI? Why manufacturing at all would need AI, is a starting point.
The index for excellence in manufacturing depends on the three pillars – Throughput or production volume, Efficiency of operation or the margins earned, Quality of product or services, and the concerns about environment – that is Safety, Health, and Emissions. Better performance means quicker decisions, which are more and more correct.
That means, better decisions help to run a business more efficiently. Decisions need an analysis of many factors affecting the particular issue, recollection and reference to similar past use cases, immediate constraints. Human intelligence can cope with the factors up to a point. Past experience meaning wisdom, is available with only some individuals. As the complexity of the issue increases, accuracy or probability of being right diminishes. This is the place where AI can step in.
AI can manipulate large numbers of factors; can have a perfect recall of all past cases, and come up with answers quickly. AI is strong in combinatorial mathematics, so having many factors to contend with is not a problem. AI is strong in pattern recognition, so it can evaluate and take reference to previous instances of the same situation, and also can recollect the actions prescribed at the previous instance.
Some areas where we can see AI stepping in manufacturing are – predictive maintenance, inventory planning and sourcing, dynamic logistics, manpower planning and skilling, quality issues. These could be the low-hanging fruit.
In a longer term we look at AI to sit by the side of every employee to augment the capability and reduce stress of his or her routine activity.
Jobs created or destroyed
Inevitably as time progresses the work humans do, and the way they go about their work will change. This statement is true irrespective of mechanisation, steam power, Electricity, and then in modern era Automation and the latest which is Artificial Intelligence.
Challenge to skills of workforce
Inevitably, jobs performed by humans will have more creative content. Repetitive and mundane jobs will be performed by machines which have different levels of intelligence starting from simple automation to Super Intelligence. So jobs available for humans will be at once more challenging and more interesting. The actual work ahead is to devise a curriculum and education system to prepare the next generation workforce to perform in such an ecosystem.
Shape of things to come
A good way to embed AI into the systems and processes of manufacturing actually needs solid foundational aspects. AI might be honoured with the designation of a Revolution when it makes sufficient inroads into the planning and implementation of manufacturing. But there are some revolutions before it which need to be honoured and given the rightful place in manufacturing.
It all started with automation. The concept of automating a machine or process arose quite sometime ago. But today, we define automation as the use of ‘Stored Program Controllers’ for performing cyclical and repetitive tasks of production machinery. This forms a basis for adding further intelligence into the factory. Summary is that a strong base of automation is a must for next steps.
Here in essence the knowledge of a human operator got encapsulated into an electronic device.
In the beginning of the new millennium, the Data Revolution came about. This focussed on the data being emanated by the automated production machinery and using this data for benefits to the factory in a measurable manner that could be evaluated in terms of money. The techniques are clubbed under a rubric ‘Data Analytics’. The analytics can recognise patterns, make predictions and give suggestions. These capabilities lay the foundation for the next step.Here in essence the decision making techniques of a supervisor get encapsulated into the analytics engine.
Presently we are in the era of AI which rests on the work of an Analytics engine. AI can augment capabilities of a manager in the plant. It is still work-in-progress but it looks like AI can acquire capabilities to run the plant just as a human manager would or could do.
What next?
The emergence of super intelligent AI will come next. To recollect from the previous episode, Artificial Super Intelligence (ASI) is a software-based system with intellectual powers beyond those of humans across a comprehensive range of categories and fields of endeavour. ASI might conceivably provide strategies and plans for companies, industries, and even countries!
In a dystopian view, ASI could run our countries and our lives making strategic decisions in a competition between ASIs. The rationale behind the decisions and their possible fallout might remain incomprehensible and unexplained to humans. The role of humans can be likened to that of domestic pets today, which are loved and well taken care of, but have no part in decision making. I repeat, this is a dystopian view.
The utopian view (utopian being opposite of dystopian) would be that the ASI has all the super competence, but remains a faithful servant to the human. To achieve this, humans have to improve themselves in terms of behaviour. They need to stop pursuing individual greed both at personal and national levels. There is a need for greater cooperation and setting common global goals, particularly in such issues as climate change, pollution and so on. Then we can focus development of AI towards tasks that benefit humanity as a whole.
Which view will prevail? I do not know, but I hope for a Utopia!
(This is the concluding part of the 3-part series on Artificial Intelligence. Part 1 appeared in the January 2024 edition, and Part 2 in February 2024)
.jpg)
PV Sivaram, Evangelist for Digital Transformation and Industrial Automation, is mentor and member of steering committee at C4i4. He retired as the Non-Executive Chairman of B&R Industrial Automation and earlier the Managing Director. He is a past President of the Automation Industries Association (AIA). After his graduation in Electronics Engineering from IIT-Madras in 1976, Sivaram began his career at BARC. He shifted to Siemens Ltd and has considerable experience in Distributed Systems, SCADA, DCS, and microcontroller applications.
Sivaram believes strongly that digitalisation and adoption of the technology and practices of Industry4.0 is essential for MSME of India. He works to bring these concepts clearer to the people for whom it is important. He believes SAMARTH UDYOG is nearer to the needs of India, and we must strike our own path to Digital Transformation. Foremost task ahead is to prepare people for living in a digital world. He is convinced that the new technologies need to be explored and driven into shop floor applications by young people. We need a set of people to work as Digital Champions in every organisation.
_________________________________________________________________________________________________
For a deeper dive into the dynamic world of Industrial Automation and Robotic Process Automation (RPA), explore our comprehensive collection of articles and news covering cutting-edge technologies, robotics, PLC programming, SCADA systems, and the latest advancements in the Industrial Automation realm. Uncover valuable insights and stay abreast of industry trends by delving into the rest of our articles on Industrial Automation and RPA at www.industrialautomationindia.in
.jpg)
.jpg)
.jpg)
