In a speech last month, President Donald J. Trump outlined three executive orders he signed to promote American dominance of AI technology. He promised, in fact, that as America is the country that started the AI race, “I’m here to declare that America is going to win it!”
To do so, his executive orders would make it easier for companies to build AI infrastructure, speed up the permitting process by doing away with any oversight and safeguards he believed to be onerous, and push the exporting of America-manufactured AI products.
This is just the latest salvo in what is now a global AI arms race, backed by billions of dollars in investment by tech companies and venture capital firms big and small here in the U.S. and across the globe. I’ve conducted academic research on responsible AI for over 35 years, I have been at the forefront of operationalizing AI — from pioneering academic research to leading enterprise adoption of analytics and AI across industries. I’m concerned about a catastrophe in the making, with giant tech firms winning the battle of the current AI arms race but most assuredly losing the war in terms of creating an utterly destructive impact on society.
The return on investment on all this money is hard to predict, other than the fact that Big Tech will become Colossal Tech the world over. Ultimately, though, success will be measured not by how much money companies and countries invest and earn but how all this AI is used and what protections will be enacted to ensure that its myriad uses are constructive rather than destructive.
For now, there are so many unanswered questions to ponder, questions that few have ventured to honestly and thoroughly answer because there is both too much unknown and for the most part the industry is totally unfettered.
This scarcity of AI skilled workers does not factor the rapid advancements in the technology in relation to the time it takes to train an individual in that technology. While it can vary significantly, depending on the complexity of the AI model, training one person can take many months. Workers don’t just have to learn new AI concepts and models, they must “learn how to learn” in a world where AI innovations are coming fast and furious. For companies, this is another significant financial investment in education and training — a totally wasteful one because there is no predictable ROI.
So, rather than focusing so much on embracing autonomy, companies must look at AI as augmentation, a collaboration between machine and human. If not, they may win the battle in the current AI arms race but they will most assuredly lose the war because the impact on society will be nothing short of catastrophic.
With such conflict between the potential for societal good and that for great harm, many wonder if there is a middle ground, and does it lay with unified global regulation of the entire AI industry. The fact is the cat is already out of the bag — every country has AI. So, I would argue that it is less about regulation and more about using AI responsibly while earning the trust of users. In other words, there still needs to be some regulation but it must be combined with common sense. It must be advanced at a pace that society finds acceptable and earns the trust of individuals and the collective whole about the short-term and long-term impacts of the technology, rather than having it shoved down their throats.
The air safety model that was adopted after World War I and that has evolved in the years since serves as an ideal corollary. Back then, as commercial air travel became a transportation option, safety measures were limited by the existing technologies, leading to many accidents that could have been prevented. But these accidents served as valuable lessons for aviation experts. In 1926 came the establishment of the Federal Aviation Administration, which was charged with improving flights both domestically and internationally. Rules were established for airways and air traffic. Licensing for pilots and maintenance technicians became mandatory, as did certification of repair stations and their crews. Manufacturing standards for air worthiness were developed. And the designs of the planes themselves were greatly enhanced with innovations such as radar systems, cabin pressurization, communication technology, and even AI itself. Today, airline travel is the safest mode of transportation.
AI could eventually be the safest mode of innovation on multiple levels and in industries across the board. But it does need to be really safe and it needs to be secure, like a plane’s black box. Only then will it be adopted by all sectors and the ROI will have real value.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
