Artificial Intelligence (September 2023)
Artificial Intelligence (AI) is a branch of computer science that aims to create intelligent systems capable of emulating human-like cognitive abilities. AI systems leverage algorithms and vast amounts of data to learn, reason and make decisions autonomously, without explicit human programming for every task. Machine learning, a key component of AI, enables these systems to improve their performance through experience.
AI encompasses various subfields, including natural language processing, computer vision, robotics and expert systems. It seeks to solve complex problems, recognize patterns and predict outcomes in diverse domains such as healthcare, finance, transportation and entertainment.
How is JPMorgan Chase using artificial intelligence?
JPMorgan Chase (JPMC) has invested heavily in AI research and development, and it is using AI in a variety of ways to enhance its customer experience, mitigate risk and make better investment decisions.
Here are some public facing examples of how JPMC is using artificial intelligence:
- Automation: AI can be used to learn patterns and automate repetitive tasks. JPMC’s COiN (COntract INtelligence) system is being used to automate tasks such as interpreting loan applications. This has freed up human employees to focus on more complex and value-added tasks. COiN is able to process 12,000 contacts annually and save approximately 360,000 hours of review for the firm's legal teams.
- Investment research: JPMC is using AI to conduct investment research. The system, embedded within the JPMorgan Markets application, can analyze large amounts of data to identify investment opportunities and provide personalized reports to bankers. This has helped JPMC make even smarter investment decisions, leading to better returns for the firm’s customers.
- Risk management: Morpheus can analyze large amounts of data to identify risks associated with trades and financial modeling data. This has helped JPMC to manage risk more effectively, providing better protection to customers and improving the firm’s own financial stability.
- Personalization: JPMorgan is using AI to personalize customer service. JPM Intelligent Assistant learns about customer preferences and needs over time. This allows JPMC to provide more personalized customer service, such as recommending products and services that the customer might be interested in or answering support related inquiries.
Is artificial intelligence the answer to all our problems?
No, AI is not a universal solution to all problems. While it excels in tasks involving data analysis, pattern recognition and automation, it falls short in addressing complex, context-sensitive and morally nuanced challenges. Human creativity, emotional intelligence and ethical judgment remain irreplaceable in fields like art, ethics and interpersonal relations. AI's effectiveness is also contingent on quality data and suitable algorithms, often requiring significant financial resources to research, train and build AI systems. Concerns about bias, privacy, legality and unintended consequences further limit its applicability, at least for the time being.
Are there any downsides to AI?
AI's potential impact on society is substantial, from enhancing productivity and efficiency to enabling breakthroughs in scientific research. AI also poses ethical concerns, however, requiring careful consideration of its implications and responsible development, and ensuring that the technology benefits humanity while respecting privacy, safety and fairness.
One of the primary ethical concerns with AI is the potential for biased decision-making. AI systems are trained on vast datasets that may contain biases present in the data, leading to discriminatory outcomes. This can perpetuate societal inequalities and impact vulnerable communities negatively. It is essential to ensure that AI models are trained on diverse and representative datasets, and developers must implement fairness and transparency measures during the model training process.
Privacy and data security are also significant ethical issues related to AI. AI systems often require extensive data collection, raising concerns about how personal information is handled and protected. Unauthorized access or misuse of sensitive data can have severe consequences for individuals and society as a whole, necessitating robust privacy safeguards and responsible data governance.
Moreover, the use of AI in autonomous systems, such as self-driving cars and weaponized drones, raises ethical dilemmas regarding accountability and responsibility. Decisions made by AI systems in critical situations may result in harm, and defining liability becomes a complex issue.
Lastly, there are concerns about the growing power and influence of large tech companies in controlling AI technologies. Ensuring open access, transparency and collaboration can help mitigate the concentration of AI power and promote a more equitable distribution of its benefits.
Addressing these ethical concerns requires collaboration between policymakers, technologists, ethicists and society at large. Developing clear guidelines, adhering to ethical principles and fostering an inclusive approach to AI development can pave the way for responsible and beneficial AI deployment in the future.
Will AI take my job?
This is a tricky and sensitive question that doesn’t have a simple answer. While artificial intelligence might be able to take over certain tasks in the next few years, the vast majority of job roles will not be directly taken over or eliminated by AI.
Jobs that involve routine, repetitive and well-defined tasks are at a higher risk of being automated. For instance, jobs like data entry, basic accounting tasks or certain manufacturing roles might be more susceptible to automation through AI and machine learning (ML). Roles that involve complex decision making or a “human touch” are at a lower risk, but these roles might still be positively impacted with advancements in AI.
Many AI applications serve as tools that complement human skills rather than replace them. For example, a complex AI model can help a team of investment bankers more accurately and efficiently analyze copious amounts of data, but the bankers themselves would still be required to review the results and make a final decision.
It’s important to note that as the field of AI & ML matures and evolves, new job roles and functions will be created. As of right now, there’s a shortage of data scientists, analysts and engineers, as well as ML engineers, all of whom are essential for the creation of AI models. Given the changing landscape of technology in general, there’s a high likelihood of new jobs we haven’t imagined yet emerging in the future as well.
What does it take to create an AI model?
In technical terms, AI works by leveraging various algorithms and models to process data, learn from it, and make intelligent decisions or predictions. Here's a slightly more detailed explanation of how a typical AI model is created:
- Data Collection: The first step in AI involves gathering and preparing the data required for training the AI model. This data can be structured (e.g., databases, spreadsheets) or unstructured (e.g., text, images, audio). The quality and quantity of the data are crucial for the success of the AI system. There’s a famous computer science adage: GIGO or “Garbage in is Garbage out”; failure to utilize a high quality data set can easily make or break any data-focused application, and that includes AI systems as well.
- Data Preprocessing: Raw data often needs to be cleaned, transformed and standardized before feeding it into the AI model. This step helps remove noise, handle missing values and ensure the data is in a suitable format for analysis.
- Feature Extraction: In many cases, AI systems require selecting relevant features or patterns from the data to focus on important aspects of the problem. Feature extraction helps reduce the dimensionality of the data and improves the efficiency of the AI model.
- Algorithm Selection: Depending on the nature of the problem and the type of data, different AI algorithms are chosen. Common AI techniques include machine learning algorithms like decision trees, support vector machines, k-nearest neighbors and neural networks.
- Model Training: In supervised learning, the AI model is trained using labeled data, where inputs are paired with corresponding outputs. During training, the algorithm adjusts its internal parameters to minimize the difference between predicted and actual outputs.
- Model Evaluation: After training, the AI model is evaluated using a separate set of data called the test set. This step helps assess the model's performance and generalization capabilities on unseen data.
- Hyperparameter Tuning: Many AI algorithms have hyperparameters, which are settings that impact the model's learning process. Tuning these hyperparameters optimizes the model's performance.
- Inference and Prediction: Once the AI model is trained and evaluated, it can be used for making predictions or decisions on new, unseen data. The model processes the input data through its learned parameters and produces an output.
- Continuous Learning and Feedback: Some AI systems can continuously learn and improve their performance through feedback loops. Feedback helps the model refine its predictions over time, making it more accurate and adaptive.
- Deployment: Finally, the trained AI model is deployed into production environments where it can interact with users, process real-time data, and provide intelligent responses or actions.
The specific technical implementation and complexity of AI systems can vary based on the AI domain (e.g., machine learning, natural language processing, computer vision) and the algorithms used. As AI technology continues to advance, researchers and developers are continuously exploring new techniques and models to push the boundaries of what AI can achieve.
In the next newsletter, we'll address Generative Artificial Intelligence, a trailblazing subset of the AI & Machine Learning field making headlines on a daily basis.
About Dan Alvarez
Dan Alvarez began at JPMorgan Chase in June 2016 as a summer technology analyst/ infrastructure engineer, and left in April 2022 as a Senior Software Engineer in Global Technology Infrastructure - Product Strategy and Site Reliability Engineering (SRE). Since May 2022, he has worked for Amazon Web Services as an Enterprise Solutions Architect.
He is also an avid guest lecturer for the City University of New York and has given lectures on artificial intelligence, cloud computing and career progression. Dan also works closely with Amazon's Skills to Jobs team and the NY Tech Alliance with the goal of creating the most diverse, equitable and accessible tech ecosystem in the world.
A graduate of Brooklyn College, he is listed as an Alumni Champion of the school and was named one of Brooklyn College's 30 Under 30. He lives in Bensonhurst, Brooklyn.