AI, Machine Learning & Data Science: How to Talk the Talk

From customer service chatbots to extremely sophisticated autopilot driving machines, artificial intelligence (AI) is undoubtedly having an impact on the world around us.

As this technology continues to become more and more present in our everyday lives, there’s a good chance you’ve thought about how you could put AI to work in your business.

But what is AI? Maybe you know a bit about what AI can do, but have yet to find an understandable definition. On top of this, when people talk about AI, other phrases pop up that sound similar and yet aren’t interchangeable – namely, machine learning and data science.

Don’t despair. We’ve got you covered. Pretty soon, you’ll be confidently chatting about AI adoption with the best of them.

Ready to get started? Here’s our quick-and-easy breakdown:

What is artificial intelligence (AI)?

AI is the science of designing machines that can complete tasks that normally require human intelligence – such as reasoning, strategizing and problem-solving.

Essentially, AI enables computers to act intelligently.

For example:

• Chatbots: Chatbots are programmed dialogue machines that simulate talking to a real human. Older Chatbot AIs had pre-programmed rules, where the designers define things the chatbot can say. Modern chatbots can learn from movie scripts - and can even be used for improv!
• Game AI: How are algorithms like AlphaZero and Stockfish able to beat some of the best chess players in the world? AI methods, including learning algorithms (AlphaZero) and expert knowledge (Stockfish), are used to create these highly intelligent game-playing systems.
• Navigation Tools: How does Google Maps find the best route to your destination? Search algorithms work to find the most efficient path between two locations without having to exhaustively evaluate every possibility.

What is machine learning (ML)?

Machine learning is a part of AI – the part that’s being adopted into industry at the highest rate. It works by taking a set of examples, building a statistical model based on those examples, and then using that model to make predictions about future samples.

Essentially, ML uses data, algorithms and computation to enable machines to learn.

For example:

• Recommender systems (e.g. Netflix or Amazon): When you’re searching for your next great binge-watch on Netflix, the app uses a recommender system, taking your viewing history to make predictions about the types of shows you would most enjoy watching next.
• Contextual web searches (e.g. Google or Bing): Did you know that if you and your friend searched the exact same thing on Google, you may get different results? Contextual web search algorithms use your search history to make predictions about the search results that will be most relevant or interesting to you.
• Speech recognition (e.g. Siri, Alexa, Cortana, or Google Assistant): How does your phone turn live sound recordings into text on your screen? Machine learning models, trained on huge datasets of prerecorded speeches and their transcriptions, are used to convert sound to text so that your virtual personal assistant can understand you.

What is data science?

Data science combines domain expertise, programming skills, math and statistics to extract meaningful insights from data. While there is some overlap between data science and machine learning, data science is often focused on explaining the data that exists, whereas machine learning is focused on predicting future data.

Essentially, data science makes sense of data.

For example:

• Descriptive statistics: When evaluating your company’s social media performance, you may look at the statistics to determine which types of posts are resonating best with your audience, and use that info to plan future posts.
• Pattern recognition: What does a company do when they have huge datasets of customer information? Data scientists use software to extract insights from data - maybe there are customers particularly open to a targeted advertising campaign, or there are some correlations between historical purchases and future purchases.
• Regression: Linear and non-linear regression can be used to identify crucial factors which drive an outcome. For example - a data scientist can help insurance providers understand which kinds of customers are the most profitable, and which ones are creating a net loss!

All together now

To recap:

• Artificial intelligence (AI) enables computers to act intelligently.
• Machine learning (ML) is a part of AI and uses data, algorithms and computation to enable machines to learn.
• Data science makes sense of data, may or may not be used in AI & ML, and can be used outside of AI & ML.

Get a crash course in artificial intelligence and machine learning in our monthly Machine Learning Foundations 1!