Two horizons lie before us. Equally attainable. The first leads to providence. The other, to ruin. We must walk the line of artificial intelligence growth well intended. A slip to either side leads to dangers. Unregulated handover of controls to the few and mighty leads to absolute corruption. The overabundance of caution and regulation leads to weak A.I. solutions that fail to come close to any tangible ROIC. To paraphrase Charles Dickens, is A.I. about to bring the best or worst of times? Both fates are equally feasible and lie directly in front of us.
Man Must Be the Arbiter of A.I.—Or the Risks Multiply
We have an A.I. tiger by the tail but must steer it gainfully. Man has to be in control. We have to ensure that the tiger exists to serve us usefully. How is this accomplished? By ensuring that ethics and values are built into this tiger from the onset. We have to tell it early on not to eat the man. Untamed tigers can be catastrophic, but A.I. harnessed and pointed in the right direction can yield tremendous growth worldwide and cure some of the planet’s biggest maladies today.
A.I. is still a child at this point. As a parent, I would never dream of sending my child into the world without instilling morals, ethics and a sense of right and wrong. While in this race for a true super-intelligence, we cannot lose sight of the fundamentals of infancy. At this baseline, we are morally obligated to code true ethics and morals into any A.I. before letting it loose in the world. It cannot simply find its own way. It will be corrupted and produce outcomes that are antithetical to what we need them to be.
Falsehoods in A.I. Hype
Another danger is that this marketing-pumped-up A.I. tiger turns out to be a mouse when it comes to returns on invested capital. Again, man has to carefully navigate through the signal-to-noise ratio, which is rather weak in the market today. We need to judge A.I.’s capabilities by holding it against humans. Ask a few questions of A.I. that you’d ask your human staff to do. The majority of the “A.I.” apps would fail in that basic enterprise version of the Turing test.
We must develop an international objective benchmarking standard on the IQ of these enterprise digital agents. Without such a standard, the enterprises are rudderless and subject to the whims of whose marketing brochure looks more compelling. As we stand now, A.I. is in danger of becoming another industry buzzword that will disappear from people’s minds as quickly as “synergy,” “disrupt” or “take this offline.” Pay attention tonight as you watch television and take note of how many commercials claim to have a product that is “A.I. powered.” What do they even mean? Do these marketers know the first thing about the true weight behind such a claim? I doubt most do.
The Danger of Self-Governance
The true danger? When you cry wolf too many times, nobody is worried when the actual wolf knocks on your door. Bad actors in the space will hide their negative intentions and plans within the fog caused by oversaturation of the words “A.I.-Powered” or simply “Artificial Intelligence,” knowing these have become white noise to the masses. There, they find freedom to be nefarious and leverage true super-intelligence for their rewards and the rest of society’s loss. When speaking of this technology, we must hold each other to a high standard.
How many people can you trust to grade their own homework? To score their own test? To write their own reviews? My answer is “nearly all.” Nearly all people will be brutally honest when engaged in self-assessment. But, the dishonest few will sully the field for us all. That is the danger in self-governance and self-regulation. The vast majority of honest actors must hold these people accountable and take them to task in front of the world.
The Role of Government
Governments have to steer A.I. purposefully and regulate minimally, focusing on the must-haves. One has to adjust the dials of regulation optimally between no regulation and stifling regulation. Over-regulation plants a seed of fear of stalling the growth of A.I. (such is the case currently in the EU) while the rest of the world can race ahead. We cannot live with compartmentalized regulations where the physical geographical boundaries have been diluted by the previous and ubiquitous nature of A.I. technologies. The line must be walked dutifully. A slip in either direction creates a future not of our wanting or need.
Chetan Dube is an A.I. pioneer. He is the founder and CEO of Quant, which develops cutting-edge digital employee technology.
