Comes now your Big Red Car with some musings on the issue of AI and not drowning the bloody dream, but I am presenting both sides before I offer a conclusion. Here goes.
Mankind has spent 5,000 years trying to create more intelligence.
Writing. Libraries. Universities. Science. Computers. The Internet. Paint by numbers.
Now we may have figured out how to manufacture intelligence itself — and right on schedule, somebody wants to regulate the living Hell out of it.
Welcome to America, baby.
Pump those brakes, Big Red Car, now
The AI brake-pumpers have five serious arguments:
1. CONTROL — Someday AI may become powerful enough that humans cannot reliably control it.
2. BAD GUYS — AI can supercharge cybercrime, fraud, deepfakes and potentially biological and military threats.
3. JOBS — It could vaporize categories of white-collar employment faster than workers can adapt.
4. CONCENTRATED POWER — A handful of companies — or governments — could control enormously powerful models, chips and computing infrastructure.
5. WE’RE TOO DAMN SLOW — AI evolves at silicon speed. Congress, courts, schools and regulators evolve somewhere between continental drift and a three-toed sloth.
Fair arguments?
Now flip the page.
Why we should step on the gas, Big Red Car?
Tons of reasons, amigo, let me give you the top five. Here we go.
1. DISCOVERY
- Cancer. Alzheimer’s. New drugs. New materials. Better batteries. Nuclear power. Engineering. Mathematics.
- AI puts millions of additional intellectual horsepower against humanity’s hardest problems.
- Slowing AI development has a cost measured not merely in dollars, but also discoveries delayed and lives not improved.
2. PRODUCTIVITY
- Give every doctor, engineer, programmer, teacher, accountant and entrepreneur an army of digital assistants.
- That’s not replacing Google Search.
- That’s potentially another industrial revolution.
3. DEMOCRATIZED BRAINS
- The rich have always been able to buy expertise.
- AI potentially gives extraordinary expertise to everybody.
- A kid with a laptop gets a tutor.
- A two-man company gets analysts and programmers.
- A researcher gets an intellectual sparring partner that never sleeps.
- Intelligence becomes abundant.
- That is the miracle.
4. EVERYTHING GETS SOME
- Steam supplied muscle.
- Electricity supplied power.
- Computers supplied calculation.
- The Internet supplied information.
- AI supplies intellectual capability.
- Medicine uses intelligence.
- Engineering uses intelligence.
- Manufacturing uses intelligence.
- Defense uses intelligence.
- Virtually everything gets some.
And then comes the argument that should cause Washington to spill its $7 latte.
5. CHINA ISN’T WAITING FOR US #
Sure, of course, America can impose an AI timeout on America. We can committee it. Some working groups that don’t do much work.
It cannot impose the same timeout on China, evil Chinese Communist bastards.
Beijing sees AI as economic infrastructure, scientific infrastructure and military infrastructure.
- Cyberwarfare.
- Intelligence analysis.
- Autonomous systems.
- Weapons development.
- Logistics.
- Manufacturing.
- Scientific discovery.
- Command and control.
Break it down, Big Red Car, give me the big picture
Now imagine that AI really does become one of the defining technologies of the twenty-first century.
America develops the lead — and then voluntarily hits the brakes while China floors it?
What could possibly go wrong?
Perhaps we can send Beijing our safety regulations so they can admire them while passing us.
This is the missing half of the AI-risk equation.
Everyone asks:
WHAT IF WE MOVE TOO FAST?
Fine.
Now answer:
WHAT IF WE MOVE TOO SLOW?
- What is the national-security risk of China having substantially better AI?
- What is the economic risk?
- What happens to American semiconductor leadership?
- Scientific leadership?
- Military superiority?
- Cybersecurity?
- Manufacturing?
- Homeland security?
Those risks belong on exactly the same damn ledger as AI safety.
Regulate crime, not brain power
Let’s take this approach, shall we?
- Fraud using AI? Prosecute it.
- Cyberattack? Hammer it.
- AI controlling a weapon? Regulate the weapon.
- Dangerous biological assistance? Lock it down.
But don’t make the fundamental mistake of concluding that because intelligence can be misused, producing more intelligence is therefore the problem.
- Automobiles kill people. We invented traffic laws. We didn’t outlaw horsepower.
- Airplanes crash. We created aviation safety. We didn’t ban flight.
- The Internet created cybercrime. We prosecute cybercriminals. We didn’t unplug the Internet.
Human civilization has spent millennia scratching and clawing its way toward more knowledge, more computation and more intelligence.
And now — incredibly — we may have invented a machine capable of producing intellectual horsepower by the megawatt.
OK, so, Big Red Car, we build it?
Hell, yes. We’re a nation of builders. We build it.
- Build the damn thing.
- Build the power plants.
- Build the chips.
- Build the data centers.
- Train the scientists.
- Secure the systems.
- Punish the abuses.
And make damn certain the United States isn’t standing at the station clutching a beautifully written 900-page AI safety regulation while the future disappears over the horizon on a Chinese train.
Bottom line it, Big Red Car, you’re getting a little strident and tedious and I just had my ears cleaned out
OK, sorry, here it is. Ears cleaned, wow! I’ll try to keep it short.
- Don’t need drown the dream at the beginning of the beginning. Let it roll.
- We need to win the race without crashing the car.
- Pedal down.
- Eyes open.
- DON’T STRANGLE THE MIRACLE.
The Big Red Car approves this message.
Now, what the Hell do YOU think?
Best answer wins a cheeseburger, onion rings, and a sweet tea at the Crystal Beer Parlor in Savannah by God Georgia.