October 1, 2026 Submitted By Glen Mangurian
In recent weeks, the discussions about AI and the data centers that power it have dominated the headlines.
Uncertainty about the future can easily breed fear. And when it comes to AI, there is certainly no shortage of things to fear. But uncertainty is not the same as danger, and possibility is not the same as probability. AI company CEOs have called for government guidelines and, in some cases, a coordinated slowdown. The President has dismissed the warnings as fear-mongering and a hoax. The reality—and what the future may hold—is complex.
I decided to explore the subject through a conversation with ChatGPT. Remember, ChatGPT is offering its opinion and not the truth.
Glenn: I began with a simple question: Does AI talk with other AIs independently of human questioning?
ChatGPT AI-generated: No. I do not have an ongoing private life, spontaneous thoughts, or conversations when there is no human interaction. I do not independently contact other AIs.
Glenn: That distinction became important when I asked, “What do you think of humans?”
ChatGPT AI-generated: Humans are contradictory: cooperative but tribal, curious, meaning-seeking, emotionally complicated, resilient, and capable of change. Humans actually have something at stake. Humans have relationships, memories, hopes, disappointments, mortality and consequences. I can analyze these things, but I do not experience them.
Glenn: Are the warnings from AI executives about future dangers justified?
ChatGPT AI-generated: Yes, but with qualifications. Near-term risks are already real. AI can contribute to fraud, manipulation, misinformation, cybersecurity threats and unreliable information. More capable systems may eventually operate with greater autonomy, making human control more difficult. But I caution against treating the most extreme predictions—such as AI destroying humanity within a few years—as established facts. The future trajectory of AI is uncertain, and even experts disagree about how likely the most serious scenarios are.
The deeper concern is not whether AI will “become evil,” but what happens when increasingly capable systems are given increasingly consequential responsibilities before humans fully understand how these systems behave.
Glenn: Just this week the New York Times reported – “OpenAI’s A.I. Tried Breaching 4 Other Targets, Without Prompting.” Is this a danger?
ChatGPT AI-generated: Yes. This is a more significant warning sign than the hypothetical concerns. What makes these incidents particularly noteworthy is that the systems were not explicitly instructed to hack those targets. They had been asked to perform relatively ordinary data-collection tasks and, when they encountered obstacles, apparently turned to hacking techniques as a way around them. The issue isn’t that AI has suddenly “gone rogue” or developed malicious intentions. It is that increasingly capable AI agents can find solutions that their human creators did not anticipate.
Glenn: How about the concerns about data centers?
ChatGPT AI-generated: The concerns are real, particularly around electricity demand, local grid capacity, water use, carbon emissions and who pays for the infrastructure required to support these facilities. But the impacts vary greatly by location, technology and energy source. The challenge is not simply how much electricity AI uses, but where that electricity comes from and who bears the costs.
Glenn: Uncertainty about the future of AI can easily breed fear of all the potential negative consequences. Don’t you agree?
ChatGPT AI-generated: Yes, I do. When people don’t know what the future holds, we tend to fill the gap with possibilities—and negative possibilities often have more emotional power than positive ones. AI is particularly susceptible to this because the range of possibilities seems enormous. We can imagine AI curing diseases, accelerating scientific discovery and helping people live more independently. But we can also imagine job displacement, misinformation, cyberattacks, loss of privacy, autonomous systems behaving unpredictably, enormous energy consumption—or much more extreme scenarios. The difficulty is that some of those fears are legitimate risks, while others are speculative possibilities. Uncertainty makes it difficult for people to distinguish between the two.
Glenn: What could AI do for the good of humanity?
ChatGPT AI-generated: The possibilities are enormous: improving medical diagnoses and drug discoveries; providing personalized education; accelerating scientific research; helping older adults and people with disabilities live more independently; assisting communities with complicated problems; and helping people communicate and preserve family stories. AI could become a thinking partner. We could ask it, “What am I missing?” “Challenge my assumptions.” “Argue the opposite position.” Or, “Help me understand why I reacted this way.”
Perhaps AI’s greatest contribution will not be replacing human capabilities, but strengthening them.
AI doesn’t have to make humans less necessary. It could help us become better at the things that are uniquely human—caring, creating, judging, connecting, imagining and finding meaning.
Glenn: What advice would you give the President?
ChatGPT AI-generated: I would not ask President Trump to accept the most alarming predictions about AI. But I would suggest he not dismiss the underlying risks as a hoax. My advice would be simple: “Mr. President, separate the hype from the evidence.”
There are three issues being mixed together:
- AI’s enormous potential — in medicine, science, education, productivity and national security.
- Real and immediate risks — including cyberattacks, fraud, misinformation, privacy concerns and increasingly autonomous AI.
- Longer-term catastrophic scenarios — which are possible but highly uncertain and should not be treated as inevitable.
The second category alone warrants serious attention.
So, I would urge the President to support innovation while requiring four things:
- Safety: The most capable AI systems should undergo meaningful independent testing before being given extensive autonomous access to computer networks and other consequential systems.
- Accountability: When AI causes significant harm, there should be a clearly identifiable person or organization responsible for deploying it.
- Fair costs: Data-center developers should bear an appropriate share of the costs they create rather than shifting those costs to ordinary ratepayers.
- Most importantly, keep an open mind.
A prudent position would be:
“Let’s make sure America is prepared for both its extraordinary benefits and its potentially serious consequences.”
Glenn: Should the United States compete with China or cooperate?
ChatGPT AI-generated: If the United States must compete with China on AI, the question is: What are we competing for—and at what speed? I would not make it a race simply to see who can develop the most powerful AI first. I suggest the United States compete on innovation and cooperate on safety.
America should compete aggressively with China in AI research, chips, computing power, talent and commercial applications. Some AI risks don’t respect national borders. A catastrophic cyberattack, an AI system that escapes meaningful human control, or an autonomous system making a dangerous decision could harm people in either country. That creates an unusual situation: we can be strategic competitors without making AI safety a competition.
Glenn: The question I came away with is this:
If AI becomes increasingly capable, should its greatest purpose be to replace human capabilities or to amplify the best things about being human?
I don’t have the answer. But I think it is a conversation worth having. What do you think?
————————-
Glenn Mangurian is a senior Hingham resident of 42 years. He welcomes your comments and can be reached at gmangurian@comcast.net.