what is the debate over AI?
In recent weeks, the debate over artificial intelligence (AI) has taken increased shape. While the computer systems application provides massive potential for innovation, there is serious question whether or not the technology will always be subject to human control. If no longer subject to human control, what are the risks to actual human survival?
The debate began taking increased shape when Jacob Coxon, a pre-training researcher at AI research companies Anthropic and OpenAI, resigned last week. He said, “Neither company is acting responsibly,” adding, “The people building AI earnestly believe that it could kill us all by the end of the decade.” Coxon believes that the artificial intelligence industry is racing to build systems that are actually incapable of being controlled.
Add to such perspective that of former OpenAI employee, Daniel Kokotajlo, who asserted, “There’s a moment in science-fiction movies when a scientist tries to warn everyone that some monster or pathogen is about to break containment. We have now arrived there in real life.”
So let’s be informed and start the conversation, seeking to understand as opposed to simply react. We begin by introducing the concept of “p(doom).” According to Intramuralist favorite, AllSides, “P(doom) is shorthand for the probability that artificial intelligence destroys the human race; estimates vary depending on who you ask, but several leaders in the AI space have put the number above zero.”
Anything with the word “doom” at the end of it sounds ripe for a crisis. So, the million dollar question, shaping the debate is: what should we make of this?
From AllSides… Dario Amodei, the CEO and founder of AI company Anthropic, said his estimation of p(doom) is between 10% and 25%. Emmett Shear, former Twitch CEO who briefly replaced Sam Altman as CEO of OpenAI, said the risk oscillates between 5% and 50%. One Anthropic researcher puts the risk at less than 10% and another at 80%.
Sounds like we don’t know exactly what the p(doom) odds are. And remember one of the Intramuralist’s mantras: we refuse to make the uncertain, certain. While certainty is helpful for both our comprehension and comfort, sometimes we have to find a way to be ok with the mucky not-knowing. Is this a crisis?
Thankfully, no less, according again to AllSides, while we attempt to ensure control over what could happen — with left-leaning voices slightly more likely than those on the right to call for increased regulation — the reality is that “the issue doesn’t fall along partisan lines.” (Thanks, God.) Multiple factors are involved such as how accepting we are of change, how much we trust big tech, feel about government intervention, existential risk, etc.
No doubt the key going forward then will be to have honest, non-politically motivated conversation, discerning what actually is good and right and true.
Interestingly, this past week far left Sen. Bernie Sanders and far right strategist Steve Bannon stood on the same stage in D.C., both espousing the same opinion in that AI is dangerous and we should stop building the data centers that power it. Friends, these are two people who agree on almost nothing. So why all of a sudden do they agree here?
This is fascinating. And remember that “fascinating” doesn’t always mean good.
As shared by Ryan Clancy of No Labels: “Here is part of the answer. Foreign governments, including China and Russia, have been running online campaigns that push the same message. Investigators at OpenAI found a Chinese company using AI chatbots to write posts blaming data centers for high electric bills, then sharing those posts while pretending to be American citizens. Researchers at the social media platform X found thousands of accounts doing something similar. Another investigation counted hundreds of stories from Chinese and Russian state media pushing the exact same idea: that data centers help the rich and hurt everyone else.”
All that to say that not every concern is a crisis, and not every crisis is real. Each government wants to be fastest and first. Hence, manufacturing a crisis has potential benefit.
All that then to also say there’s a lot we don’t know. There’s clearly more to learn and potentially be concerned about.
So let’s beware of the loudest voices, eliminate any political side-ism, be informed rather than instantly react, and be ready for continued conversation.
Respectfully…
AR
