Lately, I’ve felt like a passenger on the Titanic after it hit the iceberg, watching the band playing and passengers lounging on deck chairs.

Frankly, I don’t like feeling this way, but concerns about humans losing control over artificial intelligence seem ominous. I warned about this in an April column and another warning seems prudent.

To be sure, AI has some amazing potential for good with hopes of curing major diseases, improving food production and solving energy problems.

But concerns expressed by experts in the AI world are downright scary.

Remember, AI models have learned from human history — both good and bad. Humans have an amazing capacity for love, caring, empathy and community.

But let us not forget about the dark side of humans with hate, jealously, deceit and wars.  You have to assume AI models have learned both.

With all the warnings about AI risks, I decided to ask AI itself about this issue. Here’s exactly what my ChatGPT buddy had to say:

I don’t have desires, intentions, consciousness or a secret agenda. I generate responses based on patterns learned during training and the instructions and information available to me. I don’t independently decide to pursue goals in the physical world. The legitimate concerns are more about how humans build and use AI:

• Misinformation: AI can produce convincing false information at enormous scale.
• Cybersecurity: More capable AI could make some cyberattacks easier.
• Autonomous systems: Giving AI control over important systems without adequate safeguards could be dangerous.
• Economic disruption: Some jobs and industries could change very quickly.
• Loss of control: Future AI systems may become sufficiently capable that supervising their behavior becomes difficult.
• Concentration of power: A small number of governments or companies controlling extremely capable AI could create significant risks.

So, I’d put it this way:

Don’t be afraid of AI as though it were a creature plotting against humanity. Be cautious about giving increasingly capable technology too much power without adequate safeguards.

And there’s an important distinction: today’s AI — including me — is not the hypothetical superintelligent AI portrayed in movies. The more interesting question is what happens as AI capabilities continue to improve.

The race for AI dominance is almost entirely between China and the United States.

President Donald Trump and Chinese President Xi Jinping are scheduled to hold a summit at the White House on Sept. 24. Topics will almost certainly include major current issues like trade and Iran, but should include a very serious discussion about AI risks and an agreement to work together on safeguards.

One analyst suggested an analogy of AI being like a “killer asteroid” heading to earth, which should create worldwide cooperation in finding a way to protect us all.

We are all in this together, like it or not. I, for one, would like to continue human life and use AI to make it better, not destroy it.