Who Is Responsible?
By
Leonard Zwelling
https://www.wsj.com/opinion/humans-are-to-blame-for-ai-failures-9e0b147d?mod=opinion_lead_pos5
https://www.wsj.com/opinion/americas-must-win-battle-of-hormuz-6b24d70c?mod=opinion_lead_pos7
Extraordinary!
That’s my take on these two opinion pieces in The Wall Street Journal on September 16.
The first is by Daniel Huttenlocher, the dean of the MIT Schwarzman College of Computing. He discusses the real source of the problem that is being so widely written about at the moment, the threat of AI to human existence. He makes the comparison between when we delegate responsibility for a task to a machine vs. delegation to a person. Let’s start with a ventilator. In this case the machine will do the breathing for the patient, but the responsibility for its performance resides solely with the clinicians who prescribed it for the patient. Huttenlocher compares that to when we delegate responsibility for a task to another human, say a medical resident. Then responsibility shifts, at least in part to the resident who, unlike the machine, can actually assume responsibility.
He puts AI in between these two “delegations.” The AI can do what the resident can in terms of making decisions and changing courses of action, but AI cannot answer for any adverse outcome any more than the ventilator can. AI is a machine. All its autonomy was granted to it by a person, yet it really can accept no blame for a malfunction. Whatever it does was somehow built into it by humans.
No matter how much the CEOs of the AI companies wring their hands, they cannot foist the blame of their badly behaving machines off on anyone else, including the federal government. If Claude is a bad boy, slapping his wrists won’t help. If the AI CEOs are looking for a villain, they can look in the mirror. It is up to them to establish the ground rules and stop blaming everyone else if their products go off the rails.
In the second op-ed, Reuel Marc Gerecht and Ray Takeyh, scholars at the Foundation for Defense of Democracies and the Council on Foreign Relations, respectively, pinpoint the challenge that President Trump’s war against Iran is at the present moment. The Strait of Hormuz is at a stalemate, but the U.S. Navy’s currently successful blockade of Iran’s exports and imports cannot go on forever. Iran has demonstrated the ability to withstand a great deal of economic pain. The East-West oil pipeline in Saudi Arabia has been successfully bombed by Iranian proxies. The Iranian surrogates the Houthis’ imminent capture of a key Red Sea strait where oil exports pass has increased anxiety in the larger world. The Islamic Revolutionary Guard shows no signs of weakening its grip on Iran. The people have no voice. The Islamic Revolution is alive and well in Tehran. The U.S. and the rest of the free world are not in a good place, BECAUSE of Mr. Trump’s war.
The authors note: “The Islamic Republic –and its menace—will end only when the regime collapses. Until then, terrorism abroad and mass killings at home will persist. The nuclear-weapons program will likely rise from the rubble since nukes surely would have prevented American and Israeli attacks.”
Their conclusion? The United States must win the “Battle of Hormuz.” Or, as is the case with AI CEOs, the solution to our current dilemma in this war is to own up to our responsibility and unleash the full force of the American military so we control the Strait just as the AI CEOs must own up to their responsibilities and control the machines they have unleashed on the world. Without that American action, the mullahs and the Revolutionary Guard have won. Without the CEOs doing the right thing, we may all be in trouble.
In both cases outlined in these essays, the solution to a current problem lies in the hands of the people who started the problem, the AI bros and Donald Trump.
Both can stop looking for someone else to blame for the fact that they are in a pickle. It’s a self-made dill spear.
If the AI CEOs really want to make sure that their programs cannot destroy humanity, do a better job of understanding and building the programs before you release them. That they have built flawed machines is no one’s fault but their own.
If Mr. Trump is stuck in the Middle East, well, who’s fault is that? In January, the Strait of Hormuz was wide open and gas was well under $4 a gallon.
I refer back to my recent blog about effective leaders owning up to their mistakes. I anxiously await the time when the President of the United States and the AI CEOs say: “I’m wrong; I’m lost; I’m sorry.” The sooner they do this, the better we shall all be.
2 thoughts on “Who Is Responsible?”
I had lost track of your blogs lately and returned today to a thoughtful commentary on AI and the Iran War. I agree with accountability for both.
AI is doing some remarkable things that probably advance the efficiency of our culture and make searching information, just amazing.
The Iran War was a MISTAKE. PERIOD!!!! The Iranians are smarter and more resilient than Trump assumed (definition of ASSUME: make an ASS of U and ME). They have yet to use their capabilities to bring terror to our homeland. We should NOT underestimate their long-term plans, and we should “declare victory” and accept their terms because we have LOST this one (as we did in Vietnam, Afghanistan, and Iraq.) and need to understand that further commitment of resources and any ground troops would be DISASTROUS!!! “Annihilating” their civilian culture is a WAR CRIME!!!! GET OUR TROOPS HOME!!!
Agree, agree, agree! I use AI all the time, but I do fear that it could get out of control and come back to harm us. The AI CEOs need to man up and make sure they have control of their products. The government is of no use on this front under Trump. I guarantee you the Chinese are giving AI a lot of thought and making sure it does not disrupt their political system. I expect no progress from the Trump-Xi summit on this issue. My guess is that Chinese have already burrowed into our power grid and are waiting for just the right to use it against us.