Using AI and ML for Good

A few weeks ago I was invited to give a talk at the Sutardja Center for Entrepreneurship and Technology at UC Berkeley. It was a wonderful opportunity to speak to entrepreneurs and up-and-coming leaders about a topic that’s been top of mind for me and many tech leaders. We all agree that artificial intelligence (AI) and machine learning (ML) will be for the next decade what mobile cloud was for the last decade. If this is the case, then how do we take advantage of rapid innovations in AI and ML as a force for good??

We’ve all read dystopian novels and seen movies that show the worst-case scenarios of artificial intelligence: from HAL in 2001: A Space Odyssey to the stories in I, Robot, the scariest potentialities inherent in AI are well covered. We have fewer cultural reference points of what AI and ML could do to positively impact our world.

Maybe I am overly optimistic, but as a technologist, I believe that the advances in AI present incredible opportunities. From sustainability to global healthcare, some of the biggest issues our society faces could be improved with the application of AI and ML. Personally, I’ve been working with the team at Fable to implement cutting-edge machine learning to address mental wellness and cognitive fitness via the power of reading for 15 minutes a day. These and so many other problems can be addressed smarter, better, and faster with the help of AI and ML.

Of course, there’s a reason so many of our stories grapple with the dark side of artificial intelligence. With great power comes great responsibility, and AI is an extremely powerful tool. Anyone who wants to leverage this needs to do so carefully and conscientiously.?

Here are a few of the top challenges inherent in utilizing AI and ML well, and ultimately for good:

• Determining the right problem to solve: You can collect all the data in the world, but if you’re not sure what metrics really matter to your business and your cause, it’s just noise. “Determine the right problem” is easy to say and hard to do, especially when it comes to innovative fields. But it’s a question you have to answer if you hope to take advantage of machine learning.
• Bias in AI/ML: This issue has gotten a lot of attention in the last few years. Every algorithm is only as good as its training data, and if there are biases in your training data—which there almost certainly will be—there will be biases perpetuated in your algorithm. The most concerning aspect of this is that tech has a halo of logic surrounding it, which makes it harder to challenge its results than if they came directly from a human. There’s no easy solution to this issue, except to be on guard for instances of bias in your training data, and to remove them.?
• Combining algorithms with human intuition: No matter how great AI gets, good old-fashioned intuition will always be an important factor in making smart decisions—and, as mentioned above, in catching biases and stereotypes. Data can point you in the right direction, but it can also be misleading (especially if you haven’t quite nailed the first bullet). Never forget that AI and ML are just tools in the arsenal of human decision-making and leadership.
• Generative AI: With widespread excitement around the launches of DALL-E and ChatGPT, people are experimenting with content creation using AI at a greater scale than ever before. What are the ethics around using existing intellectual property (IP) to generate new IP? How do you ensure credit is correctly attributed? This is a tricky issue, and there are no easy answers to these questions.?
• Responsible AI — driving the “right” behaviors: This is another issue that’s been all over the news recently. Social media is a great example of smart tech being used to drive bad behaviors. From forming addictive habits to perpetuating political falsehoods and encouraging a rise in extremism, we now see the terrible human cost of algorithms used irresponsibly. If you’re using AI responsibly, your tech won’t drive “wins”—visits, engagement, sales—at any cost. It will push your users to engage healthily with your product and, ideally, the world around them. Again, with great power comes great responsibility.
• Data security: AI and ML require large amounts of data to work well. If you’re collecting and storing your users’ data, you are responsible for storing it securely. You should be actively monitoring your data security and innovating around it—because hackers will be, too.
• Data privacy: Consumer privacy has become a major concern around the world, and the US is behind many other countries on this issue. You’ll save yourself time in the future AND be doing the right thing by implementing progressive privacy standards today. The US will likely become more strict in the future, and you might as well start off the way you intend to carry on. And, again, it’s the right thing to do.
• Education: Finally, educating your teams and consumers about AI is important. There’s still a very real fear of AI taking away jobs, just like when software and automation came along a few decades ago. It’s important to address these fears head-on and to talk about the ways that AI and ML can assist humans—not replace them.

What other steps are necessary to use cutting-edge tech responsibly and as a force for good?