What Can Artificial Intelligence (AI) Do?

Now that we know what AI is, we can talk about what AI can actually do. While there are plenty of applications to consider, we have divided them up into six broad categories.

Finding a needle in a haystack, prioritizing work to maximize impact, providing early warnings or detection against imminent threats, speeding up decisions and the decision-making process, optimizing the resources that you have available to you and enabling experiments to determine the best course of action. When we think about finding a needle in a haystack, we're referring to AI's capability to sift through massive amounts of data quickly in order to identify the data point or points that we are looking for.

Instead of having to manually search through data or recognize patterns, AI can do that for us. The COVID-19 Open Research Dataset (CORD-19) currently has over 65,000 full-text articles related to COVID. This information can be critical to help scientists and medical professionals around the world treat and prevent the disease. AI can help the biomedical research community find the most relevant information in these articles.

It can quickly browse through papers and find entities like diseases, biomolecules, drugs, et cetera, and extract complex relationships between them to provide updated and life-saving information. When you need to determine how to prioritize your workload, AI can identify key tasks that provide the highest impact. For example, consultants in Philadelphia developed a model for prioritizing building inspections based on a location's distance to nearby vacant properties, distance to certain crimes, and distance to infestation reports.

They found almost 1800 sites where the probability of an unsafe outcome, such as a building collapse, exceeded 10%. This allowed the inspectors to start with the most vulnerable locations to ensure the safety of people living in those areas.

Additional benefits could include generating better daily inspection routes or providing more information to inspectors on the existing routes so they know what to focus on when they're on site. Not only can AI provide early warnings in military or cybersecurity situations, but it can also help predict diseases.

The Radiological Society of North America suggests that AI can be used to predict Alzheimer's disease years before it occurs by the identification of metabolic brain changes. Researchers had access to data from a major multi-site study focused on clinical trials to improve prevention and treatment of this disease that had more than 2100 brain images from over 1000 patients.

They achieved 100% sensitivity at detecting the disease at an average of more than six years prior to the final diagnosis. This can allow patients and doctors to work together on preventative treatment options to try to minimize the impact of the disease instead of treating it after it's been diagnosed.

When time is an important factor, AI can be used to speed up decisions by automating processes or identifying key traits that are used to making decisions. For example, when it comes to hiring, recruiting chatbots can be used to automate the communication between recruiters and candidates.

They are useful when there is a high number of applicants, ensuring that similar questions are asked of all candidates, and answering frequently asked questions effectively. JobAI is a German recruiting chatbot. Their platform offers job seekers the opportunity to contact companies, inform themselves about the organizations and apply for jobs via familiar messenger apps such as WhatsApp and Telegram.

When you are working with a limited number of resources, AI can help optimize them for maximum value and effectiveness. The Bank of New York Mellon Corporation, commonly known as BNY Mellon, developed and deployed more than 220 automated computer programs, or bots, in 2016 and 2017. These bots carried out repetitive tasks such as formatting requests for dollar funds transfers and responding to data requests from external auditors.

The bank estimated that its funds transfer bots alone are saving it over $300,000 annually. Not only that, but the bank cut down its information request response time from six to ten business days to 24 hours, providing better service to its customers as well. Last but not least, AI can use data analytics to experiment for what works and evaluate the best outcomes.

For example, in the fight against climate change, AI can help analyze large quantities of environmental data in order to identify patterns and predict future climate variability. The University of Cambridge is using AI to identify climate trends in Egypt in order to understand how climate change will affect the growth and production of cotton. This, in turn, is driving experimentation and strategy creation with the goal of climate resilience.

These strategies aim to minimize the harmful impacts of climate change on a country's agriculture while also preparing the population to make those adjustments. While there are plenty of these types of use cases, these six categories can help you think through how you might want to apply AI within your own organization.