McKinsey Predicts AI Will be as Impactful as the Steam Engine
At the Marketing AI Institute, we read dozens of articles on artificial intelligence every week to uncover the most valuable ones for our subscribers and we curate them for you here. We call it 3 Links in 3 Minutes. Enjoy!
McKinsey Predicts AI Will Contribute $13 Trillion in Global Economic Activity by 2030
According to CNBC, artificial intelligence could contribute “an additional 1.2 percent to annual gross domestic product growth for at least the next decade.” In dollars, that’s about $13 trillion in global economic activity by 2030.
To put that in perspective, this would put AI’s growth on par with that of the steam engine.
The report goes on to predict that by 2030, 70 percent of companies will have adopted some form of AI and the majority of enterprises will be using a full range of AI technology.
In terms of leaders in the space, McKinsey shared that the U.S. and China are both racing to invest heavily in AI. Beijing is one city gunning for the top spot. AI is already built into its five-year economic plan, which runs through 2020.
Being a leader in AI is key. The report predicts AI could help these countries capture “20 to 25 percent more in economic benefits compared to current levels.”
Read McKinsey’s full report, Notes from the frontier: Modeling the impact of AI on the world economy, here.
Machine Learning Use Cases
Hacker Noon published an article this week with five of the most renowned use cases for machine learning in business—and it’s brilliant. Here are two of our favorite examples.
We all know Netflix uses machine learning to suggest movies and TV shows to its users. But, are you aware of exactly how it does that?
Instead of just using machine learning to recommend shows that match the same genre of what a user watches, Netflix uses an intelligent algorithmic approach to create a true user-centric experience. To do so, Netflix processes the shows and movies you’ve watched, how far you made it through each program and how long it took you to get there. With these insights, Netflix has the power to serve up content that users can’t ignore.
Another obvious but interesting case study in machine learning comes from Facebook. You may have noticed recently that if you’re making plans with friends and family through Messenger, Facebook automatically suggest adding an event to your calendar. While it might seem creepy, Facebook’s algorithms recognize words like “today” and “tomorrow.”
But that’s not it. Facebook has machine learning and AI baked in throughout its entire platform. Another example of this is facial recognition: when Facebook recognizes a face and recommends a friend to tag.
Read more on the machine learning techniques used by Tinder, Amazon, and Uber here.
Google’s New What-If Tool
Google is back at it with AI tools. This week on their blog they announced the What-If Tool, a new feature on their TensorBoard application that “lets users analyze an ML model without writing code.”
Their reason for creating the tool is to encourage practitioners to step up their game and act as detectives when it comes to building machine learning models. In order to do so, users need to ask questions like:
- How would changes to a datapoint affect my model’s prediction?
- Does it perform differently for various groups?
- How diverse is the dataset I am testing my model on?
Getting to the root of these problems is currently inefficient, at best. It requires programmers to write “custom, one-off codes” and ostracizes non-programmers. The What-If Tool enables all practitioners to examine, evaluate and debug machine learning models.