You’ve heard a lot about artificial intelligence (A.I.) and machine learning in recent years. You know employers everywhere are willing to pay top dollar for tech pros who’ve mastered the intricacies of these technologies. But what kind of knowledge and skills do you need to acquire to leverage A.I. to its fullest?
A new report from consulting firm McKinsey, Technology Trends Outlook 2023, posits that applied A.I. involves a key cluster of underlying technologies and concepts:
- Machine learning: Machine learning is a broad category of technologies and concepts; at its heart, it relies on huge datasets to train a system to make increasingly accurate predictions.
- Computer vision: A type of machine learning that utilizes visual data (images, videos, and more) to train a system. Automated driving and other technologies rely heavily on computer vision.
- Natural-language processing: Conceptually, NLP involves a model digesting massive amounts of text and speech in order to “learn” how to speak.
- Deep reinforcement learning: A flavor of machine learning that relies on trial-and-error to train artificial neural networks.
Machine learning involves a variety of subsidiary skills around developing and teaching models to improve. Those include:
- Data management
- Model development
- Model deployment
- Live model operations
- Tools and technologies such as cloud and database technology
Generative A.I., which is the branch of A.I. that governs emerging (and increasingly popular) technologies such as the ChatGPT chatbot, also has its subsidiary software and hardware skills you can learn, including:
- Foundation models: Deep learning models that use huge datasets to train the system.
- Application layer: What the tool’s end user actually sees (such as the prompt/search bar on Google Bard).
- Integration/tooling layer: This layer manages the inputs and outputs of a generative A.I. system.
- Hardware: You need massive amounts of processing power in order to make any of this happen.
While A.I. is a “hot” skill right now, and its practitioners can earn high six-figure or even seven-figure salaries, the McKinsey report warns that A.I.’s adoption could encounter some sizable roadblocks in the years ahead, including (but certainly not limited to): lack of resources for companies to fully exploit their A.I. plans, cybersecurity and privacy concerns leading to widespread distrust of A.I., excessive regulation and compliance, and ethical considerations.
Those are all valid concerns, and how the tech industry approaches them will dictate how A.I. evolves. For technology professionals interested in A.I., a grasp of these ethical and legal issues is critical, given their potential impact on projects and deliverables.
“While A.I. adoption globally is more than double that in 2017, the proportion of organizations using A.I. has leveled off to around 50 percent to 60 percent in recent years,” the McKinsey report adds. “However, companies that have already adopted AI nearly doubled the number of capabilities they use, such as natural language generation or computer vision, from 1.9 in 2018 to 3.8 in 2022.”