AI for Humanity: An Evolution on the Brink of Revolution
By Grégory Renard
I frequently share interesting articles about artificial intelligence and deep learning on LinkedIn for my 20,000 contacts and followers. Knowing that I’m a pretty busy guy, friends ask me how I possibly have time to find and read them and whether I have others lending a hand.
I have neither the time nor any human help. Instead, I use AI to scan 70 million articles each day, resulting in a summary that tells me, “Greg, these ones are really interesting for you.”
At my company, xBrain, which is based in both Silicon Valley and France, we’re providing the same kind of augmentation for call centers. Our Satisfaction.AI system can manage more than 5,000 voice calls in the same time a human customer service agent can handle no more than half a dozen. In one day, our machines can respond to more than 80,000 conversations, more than 2,000 of them in real time.
With AI, we can program machines to help lawyers simultaneously scour tens of thousands of documents for review or help doctors instantly extract data from tens of thousands of patient records. In five to 10 years’ time, a single remote taxi driver might be able to centrally control 10 self-driving taxis at once.
While many might see this influx of AI as a threat to jobs, it is not. We will still have call centers, doctors, and lawyers. But the work of a lawyer a decade from now will not be the same as it is today. A century ago, many people had grueling manual labor jobs in places like mines and textile mills that don’t exist anymore. As we invented more sophisticated manufacturing tools and methods, they were replaced by jobs like machinery operation that required different skills and education. It’s the same today. AI will not destroy the number or availability of jobs; it will create more and new types of them.
Yes, we will have fewer traditional jobs as we think of them now. But the trade-off will be a new generation of jobs augmented by AI, using new skills that rely on a collective intelligence shared between humans and computers. The bridge between the old and new will be education (and continuing education for the rest of us).
This is our primary challenge — we must not ignore it. Executives, entrepreneurs, government leaders, and other decision makers must understand what the future with AI holds for their industries, and help shape how students and workers are trained to enter it as soon as possible.
In Sapiens: A Brief History of Humankind, Yuval Noah Harari explains our species growth through the faculty of collective intelligence. My ultimate goal is to use data science to augment — not replace — humans, and to create a new collective intelligence. I’ve dubbed it “HIAI”, the Augmented Human era. How can we balance the best of human intelligence and the best of AI?
Even more importantly, we must ensure that AI innovations are accessible to all. Any company or industry can evolve with AI. If access is concentrated in the hands of a few, societal wealth will become far more unbalanced than it already is.
This is a premise that we call AI for Humanity.
In March 2018, French President Emmanuel Macron laid out his vision for France to be a leader in AI innovation and the AI for Humanity movement. As part of the French Tech community, I contributed to the Villani Report, which unveiled France’s national AI strategy. The report examined both AI’s opportunities and its dangers. In the U.S., President Obama’s National Science and Technology Council released a similar report in 2016 on preparing for the future of AI. AI for Humanity likens AI to electricity, and, like electricity, we must take care that everyone in society reaps the benefit.
There is no reason to fear the AI solutions we are developing. That said, however, there is a real — and very scary — potential for a lack of ethics among the people creating them.
Look at autonomous weapons, which are programmed to independently search and engage targets, or Cambridge Analytica, a political data firm hired by President Trump’s 2016 election campaign that gained access to private information on more than 50 million Facebook users. You can decide to become king of the world because you have this type of technology.
A step we can take to keep this in check is a code of ethics for AI and data science experts. So, I, along with fellow data scientists Aurélie Jean and Stephane Dadian, developed the Holberton-Turing Oath. Like the Hippocratic oath taken by physicians, ours allows experts to pledge that their work will be dedicated to serving humanity with integrity. Engineers need to be reminded that they can use their knowledge to create something really good — or really bad.
We must also take care to not over-promise AI. With its potential to replicate the functions of the human mind, there are high expectations. But AI is not a sprint; it’s a long run. Achievement takes time and collaboration. We must iterate, understand, train, retrain, and iterate again.
AI is an evolution, but an evolution spawning a revolution. The birth of electricity was an evolution as well — it took a century of attempts by other scientists and inventors before Thomas Edison produced a reliable electric light bulb in 1879 and Nikola Tesla developed the first alternating current (AC) motor about a decade later — but it created a societal revolution. We’re at that same point now with AI. And if you think the web revolution was a big wave, right now, the AI revolution is a tsunami. Google CEO Sundar Pichai called it even more profound than electricity or fire.
AI is sometimes referred to as the fourth industrial revolution, but I prefer to call it the industrialization of cognitive tasks. Throughout history, we have created tools to augment our productivity, work, and knowledge. We’ve never created tools to replace them. If we break up a company’s work into small tasks, we can identify those that are cognitive and repetitive and let machines do them. And those other tasks are not cognitive or repetitive — such as those that involve empathy and intuition — can stay with humans.
While driving my Tesla electric car, I would say it is doing 80 percent of the work but I’m still doing 20 percent. That’s the middle ground that we’re looking for — the augmentation approach. We have arrived at the next revolution of creating something new to augment not only the capabilities of our bodies but the capabilities of our brains.
Mass education and inclusion is crucial. If we are not explaining to as many people as possible what we are doing with AI, a decade from now 95 percent of the population won’t understand it. We will have a new generation playing God, and that’s clearly what we don’t want.
We must embrace AI for humanity to advance our civilization. Companies, governments, and universities must integrate AI into the core of their strategies with an ethical approach. Anticipate the tsunami and don’t be afraid when it comes.
Grégory Renard is the chief AI officer and co-founder at xBrain and a prominent figure in the French Tech ecosystem. He is also a member of the NASA Frontier Development Lab’s AI Technical Committee and a Holberton School mentor. Follow him on Twitter or LinkedIn, and watch his 2015 TEDxLille talk (in French).