ChatGPT – The Most-Mentioned Technology Tool in 2023!
ChatGPT has to be one of the most popular words in the print and TV news for the year 2023. If you have not heard about ChatGPT, you have to be the odd one.
A form of generative Artificial Intelligence (AI), ChatGPT is a chatbot that has gathered more than 100 million active users in just 2 months after its release in November 2022.
GPT stands for “Generative Pre-Trained Transformer”. ChatGPT is based on a GPT-4 set of complex algorithms, trained using more than 175 billion pieces of text and image data drawn from internet websites, books, pamphlets, studies, and articles. The enormous databases that feed into GPT-4 algorithms and deep data analytics have made ChatGPT a powerful application.
ChatGPT is also a large language model (LLM), which has advanced natural language processing capabilities such that it can converse with human users directly using natural languages like English and other languages.
ChatGPT allows common people like you and me to harness ChatGPT’s amazing capabilities. Everyone who uses ChatGPT tells you how impressive and fast the chatbot is in responding to a request written in English text. The response by ChatGPT is in seconds and is in the form of English text or images. And very impressively, ChatGPT has all the answers to any of your specific requests. Its creator, OpenAI, a startup based in San Francisco, has become a household name and a huge success story.
The Negatives of ChatGPT
Industry experts have started to warn about how ChatGPT can generate false, biased, or misinformation because the databases used to train ChatGPT may contain false, biased, and misinformation.
Hallucination is the term used to describe how ChatGPT can create fake information or provide totally wrong answers to certain queries. You may want to fact-check ChatGPT’s response with a second source if the accuracy of information is important to you.
So, for now, ChatGPT is a powerful tool to help with many first-cut information gathering, article writing, presentation decks, etc. You can save a tremendous amount of time by doing so.
You can refine the first cut by refining your request or providing further input to ChatGPT. You can then put a final touch on what you get from ChatGPT.
Despite the negatives, ChatGPT is a powerful productivity tool.
More Generative AI Tools for Consumers
ChatGPT is just one of the many generative AI tools. Many other tools can generate text, images, videos, music, graphics, and other creative content.
Other than ChatGPT, ClickUp and GrammarlyGo are also generative AI writing and content creation tools. GitHub Copilot and aiXcoder are AI code-generating tools. Midjourney and Stable Diffusion are AI tools for image and artwork design. Fliki and Runway can generate videos using AI. Soundraw is AI automated music generator.
There are many more generative AI tools.
And, no doubt there will be more and more new generative AI tools to be available in the consumer market.
ChatGPT Can Listen, Speak, and See
In September 2023, OpenAI rolled out voice and image capabilities in ChatGPT. So, ChatGPT can now listen, talk, and see.
This is a key development and a huge step change that could potentially lead to even more powerful generative AI tools.
The “Talking” Autonomous Vehicle
You might have seen in a sci-fi movie how a spacecraft pilot commands a spacecraft by verbally communicating with the spacecraft computer. The spacecraft computer responds verbally, shows sophisticated information in images and graphics to the spacecraft pilot to advise on conditions and operating environment of the spacecraft, fuel adequacy for a specific space flight, etc., and initiates problem diagnosis, if encountered, and rectification actions on its own.
With generative AI tools being able to listen, talk, and see, we could be one step closer to the human commanding AI verbally. Generative AI could be the advisor, the assistant, or the copilot to a human commander.
Imagine a fully autonomous vehicle that could be a reality in the not-too-far future. There could be no need for a human driver. The term “driver” could be forever obsolete. Humans could command an autonomous vehicle by giving verbal instructions and the vehicle could execute the tasks as per the verbal instructions.
Other than autonomous vehicles, humans could command many other system networks or entities where the data-trained generative AI could be the advisors, the assistants, or the copilots.
Smart Factories That Could “Talk”
Imagine high-quality and curated factory operations and maintenance datasets could train advanced machine learning algorithms and generative AI models such that they could collaborate their powerful data analytics and prediction capabilities on massive real-time data in texts, graphics, charts, videos, etc. from factory industrial sensors, factory inputs, and other relevant sources to generate insights on factory’s immediate future performance levels.
Being able to predict the future accurately would be a strategic advantage. Image this, as an automated factory manager or owner, each morning you could receive from the factory generative AI a report containing dashboards of production outputs, level of defects, raw materials usage, energy usage, equipment conditions and performance, and completed maintenance tasks for the previous day and week. The factory generative AI could verbally highlight the salient points for your attention. You could verbally request the factory generative AI to drill down any specific information you want more details.
In addition, the data-trained factory generative AI could, based on your production targets, predict production outputs, forecast raw materials requirements for each day in the coming week using complex prediction algorithms, and list factory equipment preventive maintenance tasks for the coming week.
On your approval, the factory generative AI could then generate raw materials re-ordering schedule, which would optimise raw material inventory levels and could plan maintenance staff and tasks schedule, which would proactively prevent unplanned equipment downtime. You would expect no or minimum unplanned factory downtime. The data-trained factory generative AI could capture any early signs of equipment defects or issues.
Your smart factory would have AI-powered cameras, inspection tools, and sensors to detect scratches, dents, anomalies, and other imperfections in products the human eye might miss.
You would validate the daily production and maintenance data as new training data for the factory generative AI to continuously improve your factory production yields.
Generative AI with the ability to listen, talk, and see would make generative AI more powerful. You could be the commander. Your data-trained factory generative AI could be your advisor and assistant to help you manage your factory. Your smart factory could listen and talk to you.
Yes, a smart factory that could talk.
To further harness the power of generative AI, you would eventually build your smart “Talking” company operations which would integrate factory, warehousing, upstream and downstream supply chain operations of all your products.
Water Supply (Water Treatment Plants, Storage Tanks/Reservoirs, and Supply Pipelines) Network
Imagine advanced machine learning algorithms and generative AI that could be trained to perform data analytics and prediction on high-quality curated operations and maintenance databases and a water supply network’s historical data from sensors such as flow, pressure, temperature, pressure, chemical, water quality, pH, electrical current, power, etc., smart meters, and even CCTV surveillance video footage. Such data-trained water supply network generative AI could seek out anomalies, malfunctions, and significant changes in performance levels from real-time data and issue alarms and alerts, initiate maintenance actions, and call upon maintenance and inspection crews with specific instructions and required actions and tasks.
The smart water supply network could even send out drones on their own where needed to inspect and gather information on water supply network facilities and equipment in remote areas and could keep the water supply network operations control centre operators and managers informed on the alert and alarm status until resolved via screen displays and verbal updates.
Advanced machine learning algorithms, statistical methods, and generative AI could provide a forecast of the coming week’s daily water usage based on historical usage patterns, water production volume versus delivery volume, determine when and how long to turn on and off certain pumps and at what water levels for the water storage tanks and reservoirs, identify in advance potential water pipe leak for performing inspections and mitigation actions to avoid water waste through leaks and plan for preventive maintenance tasks for water supply network facilities and equipment.
The data-trained water supply network generative AI could verbally highlight salient points of the forecast, and operators and managers could verbally request generative AI to drill down any part of the forecast.
Water supply network generative AI could refer critical or high-impact anomalies or malfunctions, according to pre-specified criteria, to operators and managers, with proposed resolution options, for review and approval. Operators and managers could inquire about further details and all these could be done via screen displays and verbal communications.
Essentially, we could automate water supply network operations control centre operators’ tasks to enable the operations control centre operators and managers to focus time and effort on reducing water waste, improving water quality, optimising energy usage, and enhancing overall water supply efficiency. Operators and managers would no longer need to analyse complex display data and graphics on their monitoring screens by themselves. Advanced machine learning algorithms and generative AI would enable more informed decisions to protect our precious water resources.
The water supply network operations control centre operators and managers would validate the alert and alarm resolution actions and other relevant data for use as new training data for the generative AI to continuously improve the network’s performance.
Here, we could have a “Talking” water supply network, made possible by generative AI that can listen, talk, and see.
Electricity Supply (Generation/Transmission/Distribution) Network
In similar ways to a water supply network, the powerful predictive and analytical capabilities of advanced machine learning algorithms combined with generative AI’s abilities to listen, talk, and see could reshape and enhance an electricity supply network’s performance.
Imagine electricity supply network facilities and equipment performance, health, maintenance, security, safety, and weather databases, and historical data that could have been used to train machine learning algorithms and generative AI such that AI could process and anlayse enormous amounts of real-time operations and sensors data to identify and predict performance deterioration, anomalies, safety/security, and malfunctions, and issue alerts and alarms.
The data-trained electricity supply network generative AI could initiate actions to clear the alarms and alerts and send out local maintenance and inspection crews, if required, with detailed instructions to perform inspection checks and preventive maintenance tasks to clear these alerts and alarms.
Electricity supply network generative AI could keep the electricity supply network operations control centre operators and managers informed on alerts and alarm status until resolved via both screen displays and verbal communications. Operators and managers could inquire about any detail through verbal conversations.
The smart electricity supply network could schedule preventive maintenance tasks and issue work orders for maintenance and inspection crews to perform these tasks to minimise unplanned facilities and equipment downtime that leads to unexpected interruptions of the electricity supply.
Increasingly renewables such as onshore and offshore wind farms, solar farms, and remote energy storage facilities could be part of the electricity supply network generation facilities and equipment, in addition to conventional and hydropower plants. Generative AI could detect the underperformance of certain parts of solar farms and send out drones to inspect and deploy cleaning robots if the solar panels need cleaning.
By analysing large amounts of data collected from sensors and monitoring systems, machine learning algorithms and generative AI could detect the first signs of performance deteriorations or anomalies in advance in the wind turbines and initiate preventive maintenance and inspection actions within certain timeframes to prevent unplanned wind turbine downtime.
Understanding that data is key to AI application, the electricity supply network operations control centre operators and managers would validate alert and alarm resolution actions and other relevant data for use as new training data for the electricity supply network generative AI to continuously improve the network’s performance.
We could have a “Talking” electricity supply network.
Road Traffic Network
In a similar way to water and electricity supply networks, advanced machine learning algorithms and generative AI that can listen, talk, and see could transform road traffic network operations.
A road traffic network operations control centre relies on hundreds and thousands of sensors to provide real-time vehicle speeds, number of vehicles, the time gaps between 2 vehicles, presence of vehicles at traffic signals, CCTV surveillance cameras, Bluetooth beacons, etc. to monitor traffic conditions of freeways, roads, and junctions to enable smooth and continuous traffic operations.
Accidents, vehicle breakdowns, or traffic light malfunctions can disrupt smooth and continuous traffic leading to build-up of congestion in a very short time. Smart roadside cameras could play a very important role in preventing such situations.
Imagine advanced machine learning algorithms and generative AI that could be trained with historical databases including video footages of traffic conditions that could automatically alert operators with specific details of traffic issues and propose alternative options such as activating traffic police to be on-site to control traffic and dispatching nearby emergency vehicle to be on-site to assist with vehicle breakdowns. Generative AI could provide detailed reports of such incidents to the police and the emergency vehicle so they could have full information about the incidents and their tasks before arriving on-site.
The smart road traffic network could automatically adjust traffic signal timings at nearby junctions to reduce traffic build-up and automatically display relevant messages on message boards to alert vehicle drivers to avoid a specific section of the road network. The fast response would enable a speedy return of smooth and continuous traffic flow.
The data-trained road traffic network generative AI could activate the maintenance crew promptly to repair faulty traffic signals or any malfunctioned sensor or camera.
By automating traffic monitoring tasks and processes, we could improve traffic flows and performance, improve the quality of service to road users, and lower operational costs.
What else could “Talk”? What about a “Talking” city public bus transport network, or city metro and subway network, or a regional train network?
Smart Hospital Wards that Could “Talk”
In the future world of a shrinking labour workforce due to an aging population, we need our nurses and doctors to be more productive in delivering excellent and seamless patient experiences to help patients recover sooner. A smart hospital ward that connects people, data, and technology could help to achieve this objective.
Imagine patients could be given bedside tablets that could accept touchscreen or voice inputs to order food and drink, control TV, adjust room temperature and airflow setting, lighting, and window opening, and remind patients about medications or medical treatments. Imagine robots could deliver food, drinks, lodging essentials, and medicines to patients such that nurses would not be overwhelmed by many of these manual tasks.
Imagine patients could be given medical-grade wearables or smart patches that could measure real-time vital signs such as body temperature, heart rate, respiration rate, and chest expansion, detect sneezing and coughing, and detect patient falls.
Advanced machine learning and pattern recognition could detect and identify early signals of deteriorating patients such as breathing difficulties, food swallowing problems, and high fall risks.
Imagine real-time patient vital signs combined with patient medical histories and other sources of information could be integrated and presented in digital dashboards in a ward central command centre monitoring screen visible for the ward nurses and doctors to view at any time. Such centrally available information could facilitate collaborative decision-making for a patient.
Imagine all ward medical equipment could be digitally tagged such that their real-time location and service status could be available to nurses saving them time to look for medical equipment. Imagine preventive maintenance of medical equipment could be automatically scheduled with the issue of a work order, performed, and returned to the ward to enhance equipment availability and reliability.
The hospital ward central command centre generative AI could “tell” the nurses where to locate the medical equipment, and verbally warn a nurse on duty about a patient in a distressed situation or about to be in a distressed situation to enable timely and targeted intervention.
Imagine nurses and doctors could ask generative AI for a short brief of a low-care patient’s progress or detailed insights of a high-care patient before ward and patient visits.
Smart hospital wards could help consolidate hospital nurse resources, improve patient experience and well-being, streamline clinical workflows, and facilitate communication.
A Virtual Aged Care Facility That Could “Talk”?
Could it be possible to replicate smart “Talking” hospital wards in aged care?
Could it be possible for low-care aged care residents to stay in their homes and be remotely monitored with wearables and tablets (which could be used for video-conferencing where necessary) by a central aged care command centre?
Resource Productivity is What We Need
In the future world of a shrinking labour workforce due to an aging population, we need our labour resources to be more productive.
Using AI technologies to automate tasks and processes is necessary to make labour resources more productive. Operators in a water supply network operations control centre, an electricity supply network operations control centre, or a traffic network operations control centre that integrate various systems, sensors, communication networks, databases, and analytical tools into a unified platform would no longer need to analyse complex variables or screen graphics for making important decisions by themselves.
“Talking” system networks, entities, and assets would enable fewer labour resources needed to manage these system networks, entities, and assets by harnessing the powerful capabilities of machine learning algorithms and generative AI, working 24/7, in analysing enormous amounts of real-time data and providing actionable insights and verbal communications would enable the operators and relevant crews to act on the insights even more promptly.
At the same time, AI technologies would enable our precious resources such as factories, water supply networks, electricity supply networks, traffic networks, and hospital wards to be more productive by reducing unplanned downtime, reducing waste, and improving overall efficiency, notably in communications with humans.
Ultimately, this would reduce operating costs and extend the economic lives of these precious resources.
Want to read more about technology breakthrough? Check out “My Technology Breakthrough blogs”.