Machine Learningscikit-learn 1.9.1 · xgboost 3.4.1 · Python 3.12+
Dashboard
0%

AI vs ML vs DL vs data science

Artificial intelligence (AI) is the field that builds applications which do their task without human intervention; machine learning (ML) is the part of AI that learns from data, and deep learning (DL) is the part of ML that uses multi-layered neural networks.

Last updated: 05 Oct, 2026

The four names are used loosely in job titles and course names. The video sorts them with one picture, sets inside sets, and one rule: whatever the role, the end product is an AI application.

Defining an AI application

AI and AI applications · from the Complete Machine Learning in 6 Hours video · 1:23 to 4:46

Imagine the whole universe of these applications and call it AI. An AI application does its task without any human intervention: nobody has to monitor it, it makes its own decisions and performs its task. The video gives four examples:

  • Netflix. Watch action movies for a while and the recommendation module suggests more action movies. Watch comedies and it suggests comedies. It learns your behaviour without asking you anything.
  • Amazon.in. Buy an iPhone and it may recommend headphones.
  • YouTube ads. An AI engine chooses which ad you see. The goal is a business one.
  • Self-driving cars, such as Tesla's. The car drives itself based on the road ahead.

Whatever your title, machine learning developer, deep learning developer, computer vision developer, data scientist or AI engineer, the work ends in an AI application.

Placing machine learning and deep learning inside AI

Machine learning, deep learning and data science · from the Complete Machine Learning in 6 Hours video · 5:00 to 7:54

Machine learning is a subset of AI. It provides statistical tools to analyse data, visualise it, and make predictions and forecasts. The equations inside machine learning algorithms are statistical techniques, because working with data means working with statistics.

Deep learning is a subset of machine learning. In the 1950s and 60s scientists asked whether a machine could learn the way a human learns. Deep learning tries to mimic the human brain with multi-layered neural networks, and it solves very complex use cases.

Data science overlaps all three. The notes draw it as a fourth circle that cuts across AI, ML and DL. A data scientist given a business problem may solve it with machine learning or deep learning, and may first have to analyse and visualise the data. The video's own example: at Panasonic the tasks ranged from Power BI dashboards to machine learning projects to deep learning projects.

AI is the largest set; machine learning sits inside it and deep learning inside machine learning, and data science is a fourth circle that overlaps all three; Netflix, Amazon.in and a self-driving car are AI applications.

AI vs ML vs DL vs data science

TermWhat it isHow it worksExample
AIApplications that do a task without human interventionAny method: rules, ML or DLNetflix and Amazon.in recommendations, Tesla's self-driving
MLA subset of AI that learns from dataStatistical tools to analyse, visualise, predict, forecastThe classical algorithms in this course, from linear regression to clustering
DLA subset of MLMulti-layered neural networks that mimic the brainVery complex use cases, such as recognising images and speech
Data scienceA role that overlaps all threeAnalysis, visualisation, ML and DL as the problem needsPower BI, ML and DL projects at Panasonic

Where you use AI, ML and DL

  • Reading job titles. An AI engineer, a data scientist and an ML developer all build AI applications; the title says which tools they reach for first.
  • Choosing an approach. A table of a few thousand rows with numeric columns suits the classical machine learning in this course. Images, audio and long text usually need deep learning.
  • Interviews. "What is the difference between AI, ML and DL?" is a common first question, and the nested picture answers it in one sketch.
Watch out. Deep learning is not a separate field next to machine learning. It sits inside it. A neural network is a machine learning model, and it still needs data, a cost to minimise and a test set, the same ideas the linear regression part teaches.
Try it yourself
  • Sort these into the smallest circle they belong to: Netflix recommending more action movies, a model that reads handwritten digits from images, a hand-written if-else spam filter. Say why for each.
  • Name one AI application you used today and say whether it is machine learning, deep learning or neither.
  • Why does the video say data science overlaps all three circles instead of sitting inside one of them?

Little by little, you're building something great.