Artificial intelligence (AI) is an unavoidable topic. Media reports on AI are ubiquitous, sparking widespread public concern that robots will replace current jobs. However, it's crucial to recognize that AI will become increasingly prevalent now and in the future, while also acknowledging its potential to solve real-world business and social problems.
Weak artificial intelligence (also known as ANI or weak AI) refers to systems programmed to perform single tasks, such as playing chess, identifying early stages of disease in MRI scans, and autonomous driving in an environment. While these tasks vary significantly in complexity and the advanced capabilities of AI technologies, they still fall within specific operational domains.
Strong artificial intelligence (AGI or strong AI) refers to machines possessing human-like intelligence. AI is conscious, has feelings, and emotions; this is often how AI is portrayed in popular science fiction films: such as Ex Machina, Her, I, Robot, and Westworld. AI can also have more levels, such as artificial superintelligence, where robots can surpass human intelligence in multiple fields and tasks, such as a self-driving car going to a hospital to detect a patient's illness and then defeating a human in chess.
Despite the tremendous progress made in artificial intelligence research over the past few years, it may still be some time before a breakthrough in AGI is achieved; AI researchers speculate that the realization of AI could take anywhere from 10 years to over 100 years.
AI replaces tasks, rather than performs them.
There's a significant misunderstanding surrounding the tasks that automation and AI replace in jobs, and the jobs themselves. As we understand from ANI, individual tasks can be automated, but the variety and dynamic nature of the tasks that the average worker needs to perform is underestimated. McKinsey estimates that less than 5% of jobs are 100% automated. New jobs are created by automating certain recurring, single tasks to increase productivity. For example, IDC predicts that by 2023, 25% of major retailers will explore or deploy in-store robots to reduce repetitive tasks, thereby increasing worker productivity by 40%. Some job losses are possible, but this can be addressed through skills acquisition and upskilling using innovative training tools such as augmented reality.
Artificial intelligence in practice today can save lives.
Advances in complementary technologies such as cloud computing, sensors, and processors have driven cross-industry AI research, unlocking use cases and solving problems once considered insurmountable. Like most emerging technologies, artificial intelligence aims to address real-world challenges plaguing industries such as healthcare.
In the healthcare industry, there is a global shortage of medical personnel, with an estimated shortage of up to 120,000 doctors even in developed countries such as the United States by 2030.
AI can expand doctors’ expertise through emerging deep learning technologies based on artificial neural networks, easing this skills gap and serving as a complementary tool for time-constrained medical professionals.
A deep learning computer vision model called DeepGestalt, using its surface analysis framework, has been able to identify more than 215 genetic syndromes with 91% accuracy. Researchers at the University of California, San Diego, used AI to analyze structured (health records, test results) and unstructured (handwritten notes) pediatric patient data, achieving 95% accuracy in diagnosing sinus infections, 97% accuracy in diagnosing acute asthma, and 90% accuracy in diagnosing mononucleosis.
Artificial intelligence can save lives but will not put medical professionals out of work; Microsoft’s healthcare robot service is designed to streamline more direct customer service issues, while complex issues still require human intervention, thereby optimizing time and improving interactions with patients.
While other AI applications across industries aren't directly saving lives, they are still improving our safety. To save lives in various ways (94% of collisions are caused by human error, according to the National Highway Traffic Safety Administration), ANSYS is using simulation technology to train autonomous vehicles through virtual driving scenarios, thereby improving its autonomous driving AI algorithms without jeopardizing human safety. Rockwell Automation's Sherlock project is creating a safer environment by improving operational efficiency through AI modules, reducing false alarms from boilers, pumps, and coolers in industrial environments.
If AI is the brain, then data is the lifeline.
Data accessibility is a critical element for any AI deployment. The DeepGestalt computer vision system was trained using 26,000 patient cases, while a case analysis at UC San Diego utilized over 100 million data points from more than 1.3 million patient visits at a large medical center in Guangzhou, China. Deploying fine-tuned inference AI models in the real world requires massive amounts of training data. In the use case of autonomous vehicles, approximately 8 billion miles of trouble-free testing are needed to achieve performance comparable to human driving.
Through an increasing number of connected devices, Industrial Internet of Things (IIoT) platforms are contextualizing data across an organization's high-value and diverse assets. This data can then be fed into AI models to generate AI-driven analytics, either to leverage their strengths or for application in the cloud. 3D CAD models can provide annotated product datasets to train object and image recognition applications, enabling AR experiences used by service technicians. AI in the product conception phase originates from generative design simulation software, analyzing data in relation to various constraints, materials, physical properties, manufacturing processes, and design objectives.
Summarize
As these data inputs increase, AI-generated use cases continue to grow dramatically in both variety and scale. IDC predicts that spending on artificial intelligence systems will reach nearly $78 billion in 2022, far exceeding the $24 billion in 2018. The story of "co-enhancing" between humans and machines will continue to evolve rapidly; 67% of executives surveyed by PwC claimed that using both human and AI simultaneously created a stronger value proposition. The primary theme of AI will no longer be replacing people's jobs, but solving the problems facing our customers, industries, and society today.
