« Tablo Replaces Amazon FireTV Recast for Over-The Air-Viewing and Recording | Main | Artificial Intelligence Can Cure Educational Disparities for Real »

September 28, 2023

What’s More Dangerous: AI or Elon Musk?

Both have great potential.

Elon says that AI is a “civilizational risk” which requires government regulation. He also tweets “Like Gulliver, tied down by thousands of of [sic] little strings, we lose our freedom one regulation at a time.” Possible explanation of this paradox: Musk would like development in AI tied down by thousands of little strings until he and his new company xAI can catch up with OpenAI, which he helped found and lost control over.

In some contexts Elon plausibly claims leadership in artificial intelligence because of his achievements with autonomous machines including self-driving cars and the amazing rocket-boosters landing themselves for reuse on robotic barges in the ocean.  Notably he doesn’t have “civilizational” concern over this technology although millions of computer-controlled cars on the road and rockets coming down from space are better armed to do immediate harm to humankind, were they so inclined or so programmed, than disembodied chatbots running in the cloud.

Right after the Russian invasion, Elon’s Starlink helped save Ukraine by enabling the defenders to communicate even as the aggressors destroyed much of the fixed infrastructure. According to stories I’ve heard, Starlink engineers defeated Russian hackers trying to shut the links down. When the Defense Department couldn’t decide whether to pay for Starlink terminals, SpaceX decided to keep supplying them to Ukraine anyway. But later, according to Walter Isaacson’s biography, Elon decided not to allow Starlink to be used as part of a Ukrainian attack on the Russian navy in Crimea because he thought it might lead to a wider war. [He later backtracked and claimed that he really had no discretion because US trade embargos meant that he couldn’t legally activate Starlink over Russian-occupied Crimea.]

Sen. Elizabeth Warren, among others, questions whether billionaires should be making foreign policy. It’s a good question although one might ask whether foreign policy, instead, should be made by millionaire Harvard professors. Without the technology Musk is responsible for which made Starlink possible and without the billions he spent through privately owned SpaceX to make it happen, there would’ve been no decision to be made. The US has a strategic advantage because SpaceX is based here and because we are deploying a military version of the service which will be under full control of the military and not Musk, On the other hand…

Artificial intelligence is already enabling medical research which will save and improve lives. AI is a leveler, which will allow those who’ve received a poor formal education to compete with the educationally well-endowed. Yes, it can be used to write an essay which a student claims as her own. It can also write a better resume and employment letter than she might have been able to without its help and do the bureaucratic part of a job for her. AI is used both to help hackers scam and to defeat scams from hackers.

OpenAI, Google, and Facebook trained their LLMs (Large Language Models like ChatGPT) to be politically correct and as inoffensive as they could make them. Elon says that the xAI product will not be trained in political correctness. Is that a double danger to civilization or an opportunity for that LLM to make discoveries further from the beaten path?

Musk and AI both have huge potential and pose huge risks. I don’t think we should attempt to put either the genius or the genie back in the bottle. We will have to watch them both.

See also:

Artificial Intelligence Can Learn But It Can’t Think

Is AI Dangerous?

Why Artificial Intelligence Will Lead to Job Growth

Better Learn to Do Carpentry

| Comments (View)

Recent Posts

An AI Debate

I’m Gonna Be MAGA-Canceled

English is Now the Most Powerful Programming Language

A Cease Fire in Gaza Will Not Make Hamas Go Away

How Not to Control Disease


blog comments powered by Disqus
Blog powered by TypePad
Member since 01/2005