20 Things to Expect from AI in 2020 and Beyond

2020 promises to be the biggest year for AI to date. Here are 20 ways we predict AI will change work across seven industries in the years ahead.
Bret Greenstein

Artificial intelligence (AI) is already assisting and augmenting us in our day-to-day lives. But in the coming year, we’ll see it become integral to our strategic decision making at work, resulting in great value to businesses across all industries.

Here are 20 ways that we predict AI will change work across seven industries in 2020 and in the years ahead.

Healthcare

Breast cancer treatment: Geneticists are using AI to recognize patterns in cancer cells to help oncologists personalize treatments for patients and increase their relapse-free rate. Because breast cancer affects women in different ways, oncologists will soon use demographic, cellular and chromosome-level data to recommend optimal drug therapies.

Cancer detection: AI is also helping healthcare providers identify at-risk patients and assisting radiologists with early detection to reduce mortalities. It will become increasingly common for healthcare providers to apply AI algorithms to mammogram, CT, PET and MRI results.

Substance abuse treatment: Mental health therapists will use AI to create the first individualized recovery models for substance abuse victims. Employing AI and diffusion tensor imaging, neuroscientists will model brain response to persistent recovery protocols associated with high-performance recovery programs that are designed to rewire the brain’s nerve pathways. AI-driven systems can then connect addicts with the programs that will change their patterns and behaviors, reducing the risk of relapse.

Addiction treatment monitoring: Psychologist, psychiatrics and behavioral specialists are using AI to better understand addiction’s hold on people in general. Over the next year, AI will be employed in monitoring response to treatment. With the patient’s permission, professionals will receive information on their geographic location, social network activity, purchases and many other behavioral factors. Deep neural networks will better monitor patients’ behavioral patterns to determine what they need at any given moment.

Early disease detection: With diseases like sepsis, minutes count. AI is helping physicians identify and treat sepsis sooner and saving lives. As technologies like machine learning and AI become more common in medicine, physicians will use that information to more accurately diagnose and treat patients.

Child welfare: Machine learning will reduce child abuse and neglect by better allocating scarce resources to those who critically need them. AI will be applied to aggregated, anonymized data to develop personalized child-welfare models. Case discovery will leverage machine learning to analyze dark data that might otherwise go unnoticed, alerting case workers to situations that require immediate intervention.

Suicide prevention and mental health support: Suicide rates are on the rise But AI will increasingly help assess suicide risk by looking at the many factors that, when evaluated in totality, can accurately identify people who are considering suicide and intervene to save them. Chatbot apps will be used to help people with anxiety and depression as well. We will see these life-saving tools improve and become more widely available in the years ahead.

Banking and Financial Services

Check fraud: AI can notify bank employees within 70 milliseconds of a fraudulent check. It’s also very fast at spotting new forms of fraud because it’s continuously training itself to look for anomalies in new data. As fraudsters attempt new methods, AI will become increasingly critical in thwarting check fraud in 2020.

Money laundering: It’s no longer practical to identify money laundering and shady financial activities through the analysis of individual bank transactions. The first multi-bank anti-money-laundering and terrorism-financing teams are using AI-based behavioral analyses to reduce criminal activities. AI is the perfect foil for what is often a sophisticated multiple-entity ring, dispersing complex money movement transactions across multiple banks to mask criminal financing activities. With an understanding of both normal and illegal transactions, AI will analyze SWIFT and ACH records to more accurately identify convoluted laundering schemes and alert banking and law enforcement officials to unlawful activities.

Credit card fraud: Traditional methods of detecting credit card fraud relied primarily on historical transactions to predict future fraud. That’s not effective because criminals move fast and change their behavior quickly. To solve this, internal fraud teams will increasingly use AI to look at huge volumes of data to model normal behavior and identify criminal behavior. Smarter use of AI will continue driving losses down, saving financial institutions, merchants and cardholders hundreds of millions of dollars.

Pharma Protocols and Strategies

Clinical trials: AI will help identify the right patients at the right time for clinical trial studies. Specialized models will enhance patient selection by reducing population heterogeneity and identifying patients more likely to have a measurable clinical endpoint and more capable of responding to treatment. AI will test all patient demographics, diagnostic and medical therapies and recommend participants who will benefit most and the dosing and frequency for best outcomes, reducing trial times, toxicity and dropouts.

Pharmacogenomics: AI is analyzing clinical trial data sets to develop individualized treatments using biologically-inspired algorithms. AI will leverage genetic, phenotypic and adverse-reactions data, enabling the creation of new medicines for patients at a much faster rate than in the past decade.

Transportation

Trucking: Autonomous trucks are already transporting goods regularly between California and Texas. And, in December 2019, the first self-driving truck made a full coast-to-coast freight run in three days. Using 360-degree onboard cameras, LiDAR and radar systems, real-time evolutionary deep neural models will generate live 3-D models of continuously updated roads and conditions. Over the next two years, we will see more bicoastal autonomous trucks on the road, with larger-scale rollouts expected in 2022. AI freight automation will improve fuel efficiency by 10% and reduce traffic collisions.

Traffic: AI-optimized traffic signals will reduce intra-city congestion and travel time by 25% with a combination of Internet of Things (IoT) systems, 5G networks and evolving deep neural networks. Current urban traffic models will be replaced with real-time representations of traffic conditions through 5G-enabled city networks. Smart traffic lights will sync with each other to keep traffic moving as efficiently as possible and communicate with cars to improve traffic flow. This will reduce commute wait time by 40%, lower vehicle emissions by 21% and speed commute time by 26%.

Air travel: By applying AI to real-time automatic dependent surveillance data, the Federal Aviation Administration will be alerted to adverse flight behaviors that lead to midair collisions. Using real-time data from 44,000 daily flights, local weather conditions and pilot performance data, pilots will be notified of potential problems and take corrective action.

Law Enforcement

Prescription drug abuse and illegal distribution: The U.S. Department of Justice (DOJ) and Drug Enforcement Agency (DEA) will expand its use of AI to reduce the number of over-prescribers, pharmacies fulfilling large quantities of addictive narcotics and opioid “pill mills.“ Using prescription-based behavioral data, the DEA will identify good and bad distribution behaviors and alert agents on the latter. AI will better inform DOJ prosecutors so that more of the doctors, clinics or pharmacies that are over-prescribing and dispensing powerful narcotics inappropriately or for non-medical reasons are brought to justice.

Helping local police: Today’s local or federal law enforcement agencies use drones, facial recognition, public records and social networks to track people and gather information on suspicious behavior. While these AI capabilities are augmenting decision making, they are disconnected. In the future, these kinds of systems will be used on a larger scale to optimize policing strategy to ensure that the right amount of law-enforcement resources are present.

Setting bail: Bail-setting is often maligned as inconsistent and unfair. Some U.S. courts have started using AI-based systems to recommend bail based on region, type of crime and criminal history. We will likely see more courts adopt such systems, but the challenge will be to continually detect and effectively address systemic biases.

Insurance

Premiums, risk and portfolios: Property and casualty insurance carriers will develop highly customized coverage policies reflecting real-world economic, environmental and political changes. Using real-time data feeds from 5G networked environmental sensors, AI will enable the micro-predictions necessary for accurate on-demand insurance coverage. Improved actuarial models will reduce premiums and improve risk ratios.

Retail

Shopping trends: In 2020 and beyond, shoppers are expected to continue their shift away from brick-and-mortar and toward online purchases. But, however they choose to shop, retailers will use the predictive capabilities offered by machine learning and evolutionary computation to inform marketing strategies and anticipate their needs. Customers will be expecting an experiential retailing experience from businesses in 2020 and beyond, and AI will help bring this into the mainstream.

2020 promises to be the biggest year for AI to date. AI-enabled intelligence will continue to improve our quality of life as well as business profitability, growth and outcomes.

Credit: cognizant