Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Step 1: Get clear on your goals
When 2017 headlines declared the retirement of a worldwide Go champion thanks to DeepMind’s artificial intelligence (AI), AlphaGo, public sentiment swung towards aversion and fear. Movies and sci-fi literature continue to herald the end of our race and jobs–rendering us useless. It’s a good story, and it sells. Despite AI amassing popularity and applications in various industries, some skepticism remains. What most of us get wrong is that AI can markedly improve human decision-making in just about anything. All it takes is choosing to understand how AI reasons.
AlphaGo demonstrates AI’s potential as a strategic tool
The better story is that we can use AI as a tool to enhance our capabilities. How we respond to AI determines its value to humans. It’s instinctive to fear what we do not understand. If we can look past defeat, we can benefit from reframing AI’s mastery of Go as a learning opportunity. DeepMind developed AlphaGo Zero, which defeated its predecessor, AlphaGo, in 100 consecutive games. How did it do that? Through endless rounds of self-play over three days. This is a stark difference from AlphaGo’s training, which consisted of learning massive volumes of human knowledge on Go. AlphaGo Zero trained itself to win. The good news is that AI can help us train to win, too. AlphaGo Teach is an interactive program that shows Go players potential moves the AlphaGo AI sees and the winning probabilities of each move. A 2021 study by Shin et al from Yale University and the University of South Carolina found that people learned to make better decisions–ones that AI would make–after understanding how AI reasons.
AI learns from us to make rational decisions
Humans trained computers to do a ton of things, like drive cars, recommend purchases based on our preferences, and recognize images. At Thinking Machines, we’ve trained AIs to reconcile customer data, migrate a telco's network, and even monitor rice productivity in Cambodia. We did that by training them on learning algorithms. Learning algorithms are vast amounts of data that AI processes and analyzes to identify patterns and relationships. Based on that, it creates rules that help it perform tasks. This, in a nutshell, is how machine learning works. That is precisely how AlphaGo learned to play Go–by analyzing millions of game moves that experts and champions have played over decades. It looked for the best plays and the patterns across them. Then, it emulated and improved those plays to win its games. Imagine having that information at your fingertips in time for your next Go match (with a human). AI learns from us by example, but that doesn’t mean it can overtake us in every regard. We remain the best candidates for handling the minutiae.
AI and humans are better together
AlphaGo Zero and AlphaGo Teach underscore AI’s value as a tool to supercharge humans. One example is DALL-E, a video and photo editing assistant. DALL-E can generate photos of subjects you’ve never seen before, like a monkey doing taxes while wearing a funny hat. It does this through a feature called inpainting. It can also augment images by changing a subject, like a dog to a cat, or create different iterations of a single painting. We can realize our ideas on canvas, no longer bound by limitations in our artistic skills. DALL-E’s capabilities show that AI can extend our creativity to new heights.
AI-generated images using DALL-E.
Step 1: Get clear on your goals
That’s fun, but let’s talk business. You can boost how you support and optimize your team’s performance, or meet your client's needs using AI. Companies like Taobao and Vodafone have deployed AI for sophisticated personal recommendations for customer upselling and identifying customers who require assistance before they even call. AI works for the social sector, too. We’ve worked with international nonprofit iMMAP to identify new informal settlements across Colombia to expand humanitarian aid. Integrating it into your business can raise your value by 5 times compared to replacing a human workforce completely with AI. The key difference is nurturing a culture of collaboration between your people and AI. As Shervin Khodabandeh highlights in his TED talk, technology alone is not a silver bullet for any problem; it requires the imagination and skills of humans. If we want to tap into the benefits of AI, we need to be proactive and involved, not reactive and afraid. Humans synergizing with AI isn’t an unlikely pairing. It’s a great one. This post is part of a series! Stay tuned for our next blog, which will tackle change management and leading AI innovation in your organization. In the meantime, check out our other posts on artificial intelligence and machine learning.