Understanding AI Models, Terms, and Policy

Artificial intelligence (AI) is transforming everything from health care to transportation, and policymakers across the United States are racing to keep up. But what exactly is an AI model? How do different types of AI work? And what are lawmakers doing about it? This guide explains key AI terms and explores current debates over AI governance.

What is an AI Model?

An AI model is a computer program trained to perform a specific task by analyzing large amounts of data. For example, an AI model might predict weather patterns, identify objects in photos, or generate text. The “model” is essentially the set of rules the computer learns from data, allowing it to make predictions or decisions without being explicitly programmed for each case.

What are the categories of AI models?

There are many types of AI models, each designed for different purposes. Some of the most common categories of artificial intelligence include:

• Machine Learning (ML): Algorithms learn patterns from data to make predictions or classifications.
• Deep Learning: A subset of ML that uses multi-layered neural networks for complex tasks like language translation or image recognition.
• Generative AI: Creates new content, such as images, music, or text, based on training data.
• Expert Systems: Uses a set of predefined rules to make decisions in specialized fields.

When comparing ML vs deep learning, the main difference is complexity: deep learning models require more data and computing power but can handle more sophisticated problems.

Artificial Intelligence Classification

Experts often group AI by capability:

• Narrow AI: Focused on a single task (such as a chatbot or recommendation system).
• General AI: Hypothetical systems that could perform any intellectual task a human can do.
• Superintelligent AI: A speculative future stage where AI surpasses human intelligence.

The race to achieve Artificial General Intelligence (AGI) is one of the most talked about topics in this field.

AI Governance and Regulation

As AI becomes more powerful, governments are weighing how to regulate it. In the U.S., American artificial intelligence policy includes initiatives like the 2023 AI Executive Order, which set federal standards for transparency, safety testing, and civil rights protections. This was later rolled back by President Trump, who is in favor of less stringent AI regulation. More recently, the Trump Administration revealed its AI Action Plan, which aims to promote the growth of American AI technology by further reducing regulation and speeding up infrastructure permitting.

States are also taking action, in some cases regulating the ways AI systems are used in hiring and law enforcement. New Hampshire has moved to ban the fraudulent use of “deepfakes” and regulate how state agencies can use AI technology.

• AI policy areas under discussion broadly include:
• Data privacy
• Bias and fairness
• Intellectual property rights
• Safety and accountability
• National security

The federal and state governments are working to support the development of artificial intelligence technology while creating safeguards to protect the public. To learn more about some of the proposed regulations, read our 2024 article about legislative action on AI.