Image of woman’s face with technology superimposed
Canva
AI is everywhere. The advances in AI have benefited virtually every commercial industry – but there is as much (or more!) hype about AI as there is real AI. In the midst of this are the terms – AI, Data Science, Machine Learning, Deep Learning, etc. – that are adding to the confusion. In this post, I offer perspective on two of these terms – AI and Data Science, and what they (commonly) mean relative to each other.
What is AI?
Artificial Intelligence (AI) is an umbrella term for any technology where a computer program is attempting tasks that come naturally to the human brain. Skills such as understanding written language, detecting speech, recognizing objects from images, and making plans to optimize time, are all examples of intelligence that humans display every day. Most are learned by our brains naturally as we grow and interact with the world around us, and are then refined and advanced by formal learning.
These tasks come naturally to humans but are quite challenging for computers. Computer algorithms (ways to structure programs) that can learn and perform these tasks are usually classified as AI.
What is Data Science?
In the same way that AI is an umbrella term for intelligence, Data Science is an umbrella term for insights from data. Data Science is a set of methods and practices for gathering insights (information, learnings, etc.) from data. The data can be anything (stock prices, voice recordings, sensor data from rainfall meters, satellite images, etc.). Data Science can include processing the data, performing statistical analysis of the data, presenting the data in ways that others can understand (called data storytelling), and so on. Sometimes these analyses are simple (like average rainfall). Sometimes they are far more complex. But it is all data science.
Does AI need Data Science?
Often, yes. Before a computer program tries to learn from the data, it is often helpful for a human (or data analysis program) to study the data. Data Scientists often clean the data, extract out important elements, and feed these to an AI to learn further from. This intervention often helps AIs learn better because the AI can focus on selected subsets of the data, thereby improving the learning process.
However, these days the most advanced AIs are able to sift through large data volumes that have undergone minimal to no data pre-processing. There is also automated software that can help clean and pre-process/select data for the AI. As such, some advanced AIs do not necessarily require classic data science.
Does Data Science need AI?
Sometimes. Data Science can be used by itself to understand, explain and communicate insights about data. For example, if rainfall data is analyzed to see if average rainfall exhibits an increasing or decreasing trend, this can be done with statistical analysis that does not require an advanced AI. However, it is possible to use AI to learn insights from data that are not visible with simple data science techniques. This is particularly true with rich data types (such as video), or when data volumes are particularly large.
Which is better?
Sometimes these two terms appear to be in conflict or competition. This is not the case. The field of data and machine intelligence is vast and involves everything from understanding data to helping computers learn from the data and solve problems automatically using their learnings. Both Data Science and AI are critical for businesses and have a complementary relationship. In the future, we can expect a seamless relationship between the two – and no need to pick one over the other.
Check out these additional resources if you would are new to AI and want to learn more, or new to Data Science and want to learn more.
