Augmenting high value industries with Artificial Intelligence

Augmenting high value industries with Artificial Intelligence

Artificial Intelligence (AI) will have an extraordinary impact on many industries throughout Ontario. The predictive capacity of AI will be especially impactful to high value added industries such as healthcare, finance, and manufacturing where new business models and efficiencies will be unlocked. The benefits are expected to be felt by both businesses and also consumers.

Healthcare

Medical professionals will benefit from AI assistance to read medical images, collate relevant medical records, and propose diagnoses and treatment options based on identifying complex underlying patterns. In 2016, a cancer diagnostic AI achieved a 7.5 percent error rating, while a highly trained pathologist reviewing the same images had a 3.5 percent error rating. But in combining the predictive capacity of the AI and the pathologist, the error rate dropped to 0.5 percent, a seven-fold improvement over the human alone. Efficiency and accuracy improvements will be among the primary drivers of productivity enhancement among complementary occupations.

While the example above demonstrates the power of AI in conjunction with a medical professional, there are advancements being made in standalone diagnostics. A project out of New York’s Icahn School of Medicine at Mount Sinai has linked 12 years of electronic health records covering 700,000 patients. By utilizing unsupervised machine learning techniques, a diagnostic model that appears to drastically outperform medical doctors’ traditional diagnostic techniques was constructed. Diabetes, schizophrenia, and select types of cancers were particular standouts. The ability to diagnose schizophrenia was of particular interest, as the AI possessed a strong ability to predict its presence in the future, while failing to reveal how it came to the conclusion. What it does demonstrate is the opportunity of increasingly personalized medicine powered by AI, and based upon a patient’s digital medical records.

Finance

In the field of finance, AI is improving the efficiency of large firms. Financial institutions already employ powerful pattern recognition algorithms to monitor and detect fraudulent transactions, money laundering, and terrorist financing. The UN Office on Drugs and Crime estimates that $1.0 trillion to $2.5 trillion dollars are laundered globally each year.[1] Efforts to contain global money laundering were estimated to cost $10.3 billion in 2017, with a significant portion driven by false-positives, accounting for 75 to 90 percent of flagged transactions.[2] Previously this task was mostly conducted by humans, who are inherently less skilled at recognizing complex patterns even when assisted by data processing tools. AI has the potential to improve pattern recognition and drive overall costs down.

Manufacturing

AI and machine learning also have applications in manufacturing. Currently, many industrial applications of robotics depend upon a carefully constructed set of instructions and algorithms to guide their actions. Allowing trained AI to control robotic entities and find more efficient solutions for a task is an exciting new frontier. Robotics managed by AI are unlikely to directly replace workers, but will further enhance an already automated process.

Artificial intelligence will have applications in businesses’ ongoing development cycle, helping product designers and engineers reach an enhanced solution in a shorter time. AI can also highlight design choices that cause issues in the manufacturing process.

Once a factory line is operational, AI applications can improve monitoring and predictive maintenance. By utilizing sophisticated sensors, to measure visual, temperature, vibration, or audio patterns, production machinery can be evaluated by centralized AI, and predict mechanical failures before they happen. Unscheduled downtime on production lines can be very costly for large-scale and high-value production; the predictive power of AI and machine learning in tasks where patterns in data are not easily detected by a human worker cannot be overstated.

To learn more about artificial intelligence and its future impact on Ontario’s economy see: From prediction to reality Ontario’s AI opportunity.

Written by Chris Mack

Photo credit: Mike Ellis


[1] Note: UNODC reports US$800 billion to US$2 trillion, which is converted by OECD 2017 PPP to C$1.0 to 2.5 trillion.

[2] Note: Wealth Insight reports US$8.2 billion, which is converted by OECD 2017 PPP to C$10.3 billion.

Category: Economic Progress, Economy, Innovation, Productivity, Technology