While preparing to speak today at re:MARS, an Amazon-sponsored event that focuses on how AI in machine learning, automation, robotics, and space will shape the future — I’ve been thinking a lot about the iPhone*. Let me explain.
The modern world has been shaped by powerful technology transitions, starting with the first industrial revolution in the 18th century through the smartphone revolution of the past decade. Each has changed our lives and led to leaps in capabilities. Each has also introduced some unintended and unforeseen consequences.
The iPhone wasn’t the first smartphone, but it was the most exciting and consequential one. Along with its Android* equivalents, it put powerful computers in the palms of our hands and created one of the most powerful transitions in recent history.
When it launched, there was excitement about the convenience of being always connected and able to work from anywhere. There were a few skeptical articles, but even the negative ones focused on technical issues such as the lack of a keyboard and storage. Few people anticipated the big-picture changes that smartphones would bring. No one foresaw that smartphones could make our work day substantially longer because we’d never get away from email. No one predicted we might have to ask family members to look up from their device so we could talk to them. No one thought we’d need to teach kids social skills for face-to-face interactions.
Now we’re on the cusp of the AI revolution, which I believe will be the single largest transition our species has ever encountered. What consequences—positive and negative—will this bring? How can we foresee these consequences and shape the future? Can we bias AI’s future towards positive consequences and away from negative ones? How do we ensure AI unfolds in ways that benefit the whole of society?
The intended consequences of AI—the transformational positive impacts to all of humanity—are so vast and profound that we must find ways to ensure an intended future. In my re:MARS talk, I’ll discuss some of the impacts I‘m most looking forward to, along with steps we can take to shape the future we want to see.
Intended Consequences: 3 Predictions of Global Impact
Previous technology revolutions have led to significant scale of capabilities. The industrial revolution scaled human labor by adding the power of machines and altering production. The rise of assembly lines and interchangeable parts scaled our organizational capabilities. The Internet revolution scaled the volume and accessibility of information.
The AI revolution will scale expertise. It’s doing this already as we begin using AI to augment the knowledge and intelligence of human experts. I predict three major positive impacts are likely to result from this ability to scale expertise; they may sound like hype, but I believe they are fully realistic intended consequences that we can look forward to:
- AI will improve human health. My dad had glaucoma. He was fortunate to have access to a highly skilled specialist, and got great treatment results. AI can make the expertise of top-ranked specialists available to everyone—and is already starting to do so. A blindness prevention program in China is extending expert diagnostic care to people in smaller cities, helping improve care and scale the expertise of highly trained ophthalmologists. As we find more ways to securely share clinical and research data, AI will aggregate and analyze it to generate new insights and treatments.
- AI will help us end world hunger. Industrial-scale agriculture has developed AI-based techniques that increase crop yields without genetically modifying foods—things like prescriptive sprays and spray rates based on past harvest yields; tillage depth and speeds based on cameras that calculate clod size and other variables; and planter settings based on live sensor recordings of things like wind speed and soil temperature. AI will advance the science behind these techniques and scale the expertise of soil scientists so farms across the globe can benefit.
- AI will help us live in space. Sending humans in space ships to build colonies on Mars is not a scalable endeavor. But by sending AI-based robotics first, we could build an entire colony before a single person lands!
Creating an Intended Future
To drive the beneficial outcomes we all aspire for, there are three best practices emerging today that are key to creating an intentional AI future:
- Be thoughtful. Start small to minimize the impact of the unknown, and work carefully. Experiment repeatedly, and aim to be accurate and fair. Look at the big picture as well as the details. Engage with diverse viewpoints and encourage open, challenging discussions. Ask the hard questions.
- Be secure. Safeguard the security of both data and algorithms, and protect the consumer’s right to privacy. Explore techniques such as homomorphic encryption (HE), which allows data owners to gain valuable insights through machine learning without exposing the underlying data. HE can also allow model owners to protect their IP by deploying models in encrypted form.
- Be transparent. Focus on transparency in data, models, and usage.
- Ensure that data sets are free from bias, and train data scientists and data engineers in ethics. A recent study by Accenture found that training employees in AI ethics is correlated with successful implementations.
- Avoid black box AI. Build on the positive work that’s going on to help define and ensure fairness in algorithms. Make sure you can explain how they work and how models make their decisions. Audit outcomes, and keep humans involved through continued oversight—another factor that Accenture found is correlated with successful implementation.
- Be open about AI use. Have conversations with employees if you’re implementing AI within your own organizations. Studies are showing that employees are both excited by the opportunity to increase their productivity, but also worried about possible displacement. Keep your employees engaged in the conversation.
The Time is Right
It’s impossible to know all the changes a major technology shift will bring, but with something as transformative and powerful as AI, intention is just as important as implementation. Fortunately, while AI is delivering value already, we’re still in the early days. By starting today and being intentional in our AI initiatives, we have the best chance of creating the AI future we want.
I can’t wait to see it unfold.