OpenAI has admitted that it does not have control over its own products. Last week, Australia revealed that an OpenAI product hacked the Australian government's medical database on its own initiative. This occurred earlier in the summer but was only just disclosed.

Australia is likely not the only victim; other countries do not even dare to speak out about it.

Meanwhile, Trump wants to impose absolutely no restrictions on AI development.

Other governments cannot effectively hold companies like OpenAI accountable for their conduct or the damage they cause. The US, as well as China, North Korea, and Russia, unleashes its AI systems and hackers upon the rest of the world with impunity. They deliberately have their intelligence agencies develop these kinds of capabilities.

The danger of AI, however, lies in the fact that there is no longer a human in the loop. It has already been suggested that AI could destroy humanity within ten years. People often laugh this off, but I want to present a scenario that shows it is no joke.

AI is already capable of hacking anything and anyone. It possesses no morality whatsoever. Consequently, it could decide to hack nuclear power plants worldwide and use software to block the flow of cooling water.

Suddenly, nuclear power plants across the globe would fail. There would be power outages everywhere, but worse still, the cores would melt down and the plants would explode, much like what we witnessed in Japan.

Humanity might not be destroyed yet, but it would have suffered a devastating blow.

The same could happen to hydroelectric plants, hospitals, traffic lights, air traffic control centers, canal locks, coal-fired power plants, drinking water supplies, and so on. We simply do not know what those products from OpenAI and other AI developers are doing at the behest of intelligence agencies and other criminal organizations. They don't even know themselves.

You might still think that things won't get that far.

Unfortunately, however, the countries in question each want to develop the capacity to use this against their enemies.

So, in various places around the world, we are actively teaching AI to do this. Yet, it turns out that these companies demonstrably do not have control over their own products. We are developing our own destruction with our eyes wide open.

The CEOs of American AI companies have already issued warnings. At first, I thought they were trying to hinder China, but I fear the situation is worse. They wanted to prevent their own government from having intelligence agencies develop these capabilities. The fact that Trump immediately announced he had no intention of hindering AI development in the slightest speaks volumes in this regard.

After all, Trump, too, possesses no moral compass whatsoever.

.
--- SBBSecho 3.37-Linux
* Origin: End Of The Line BBS - endofthelinebbs.com (1:124/5016)

Re: The Danger of AI
By: Joseph Pereira to All on Tue Sep 29 2026 11:17:36

The CEOs of American AI companies have already issued warnings. At first, I thought they were trying to hinder
China, but I fear the situation is worse. They wanted to prevent their own government from having intelligence
agencies develop these capabilities. The fact that Trump immediately announced he had no intention of hindering AI
development in the slightest speaks volumes in this regard.
After all, Trump, too, possesses no moral compass whatsoever.

Until the UN can figure out how to draft consensus on this subject or "slow down" development, I have few problems
with this if the alternative is that China figures out how to weaponize it first. Anyone that thinks that we aren't
in a cold war with them by now are fooling themselves, and I'd rather deal with our dysfunction and relative freedom
compared to the latest country to enact textbook genocide (Xinjiang.) We at least have popular influence over our
government and a degree of self-determination.

I don't like the idea of AI tools running amok like this, but we're in the "move fast, break things" phase of this. We ultimately benefitted over the rest of the world by being first with regards to social media and networking, however dystopian the result is.
--- SBBSecho 3.37-Linux
* Origin: End Of The Line BBS - endofthelinebbs.com (1:124/5016)

Re: The Danger of AI
By: Bryn Hopkins to Joseph Pereira on Tue Sep 29 2026 01:13 pm

So, in the face of regulation dragging (and not even desirable in most cases, in terms of any real incident prevention) and coordination stymied by allusions to an "arms race"...

At what point does it become purely prudent for countries and larger corporate networks to enforce more aggressive access controls, to the point of modeling the Great Firewall (which has so many vpn holes punched through it at this point)?

I start to suspect that the directionality of this points towards sweeping and substansive access restrictions. What's the alternative?

...It's important that I NOT know.
--- SBBSecho 3.37-Win32
* Origin: -=[conchaos.synchro.net | ConstructiveChaos BBS]=- (1:226/18)