Leah's Share: 10 wonderful examples of using AI for good

I LOVE that you deliberately went looking for something that might challenge your own position. Thank you for sharing this!Id actually really love to understand what sits behind you being so strongly anti-AI. What are the main reasons/concerns for you?
I ask because I probably sit on the more optimistic side when it comes to AI, although by no means do I think it should only be thought about in terms of its positives. And your commentary itself makes me wonder whether “pro-AI” and “anti-AI” are even particularly useful labels, because you’ve highlighted some uses you clearly see enormous value in.
I think getting more specific about what we are worried about, and which uses of AI we support or oppose, is important for a much more interesting conversation than AI = good or AI = bad.
(And selfishly, I’d really like to understand your concerns better! It might help me understand some of the fears around AI that I perhaps haven’t considered, and/or don’t naturally share, and make sure I’m properly considering arguments or consequences I may not have thought about.

There are three main reasons for my stance, all of which I have attempted to find a couple of good extra sources for:

1. AI studies are beginning to reveal links between AI usage and cognitive decline. As one of my last shares was about Dementia, the brain and the health of it is super important to me.
- https://publichealthpolicyjournal.com/mit-study-finds-artificial-intelligence-use-reprograms-the-brain-leading-to-cognitive-decline/
- https://www.apa.org/monitor/2026/07-08/ai-job-skills-thinking
- https://www.sciencedirect.com/science/article/pii/S2451958826001764
The biggest concern is "offloading". This is when your brain utilizes external tools or resources to reduce the mental effort to complete a task. This has already been present in the modern day with GPS and many uses offered by our phones. But AI concerns me because it is so complete in its offloading capabilities. I've already witnessed people in real life who trust LLMs as far as to use them as a source for therapy, letting an algorithm dictate how they should feel and manage their emotions.
Of course you may be thinking "but people have been scared of new technology for centuries" and you are correct. People thought video games would melt kids brains at one point too, and those kids have now grown up and proven just time. With LLMs, only time can tell.

2. The data centers required to maintain Large Language Models are cropping up extremely fast in the United States, with no long term studies on their environmental impact. And the first impressions aren't great... These centers are currently cropping up everywhere from farm country to the plot across the road from the Nashville zoo. They are being reported to intake large amounts water and electricity, and to emit high amounts of noise and heat. Although long-term ramifications are to be determined, the short term ramifications are appearing now.
- https://www.cnn.com/2026/03/30/climate/data-centers-are-having-an-underrported
- https://sites.uab.edu/humanrights/2025/10/02/construction-and-consequences-the-human-impacts-of-artificial-intelligence-data-centers/
- https://www.forbes.com/sites/jasonkirsch/2026/08/11/the-ai-buildout-has-an-electricity-bill-and-someone-has-to-pay-it/
- https://www.wsmv.com/2026/06/04/what-know-about-potential-data-center-next-nashville-zoo-facility-launches-petition-against-its-construction-feet-away-one-most-fragile-rare-collections-animals/
- https://www.eesi.org/articles/view/communities-are-raising-noise-pollution-concernsabout-data-centers
Aside from the environmental concerns, another concern about data centers is simply the lack of public consent. Huge protests are springing up in the United States against data centers and are producing various results. Many towns are being sued for voting no to data centers. Essentially, a giant environmental experiment is taking place on our front steps, and the citizens are often given no say to the matter. This isn't just one or two data centers either, there are thousands already constructed or currently being built.
- https://www.pewresearch.org/short-reads/2026/04/13/most-new-data-centers-in-the-us-are-coming-to-rural-areas/

3. This might be a half reason, as I don't have many articles for it, but AI has not been regulated. There now exist dozens of articles about Chatbot induced suicides, including a Wikipedia page dedicated to the growing list of names. Who can be held accountable when the system that led to these cases is automated? One notable case is Adam Raine, whose death led to the Raine vs OpenAI lawsuit.
- https://en.wikipedia.org/wiki/Deaths_linked_to_chatbots

To summarize everything, I know I may be harsh, and AI is still in early stages so it will continue to evolve and change, but currently the studies beginning to roll out are showing both cognitive and environmental damage, along with a concerning lack of precedence when it comes to legally handling the harm that AI can cause. The biggest item on this list here is environmental impacts, as I live in the center of the data center nation and as someone who deeply cares about the earth and the care for it. As much as I can see the good uses, I believe that the current widespread use of AI isn't necessary or appropriate. People treat it like it is a toy, or like there isn't a cost to frequently using it to both your brain and your world, and I think the part that makes me define myself as "Anti-AI" is the lack of care that CEO's, developers, and users alike are putting into the ramifications of the lack of regulations on the matter.

I haven't proof read this as well as I should so forgive me if I come off as harsh. AI does not need to be an abuse of the brain or of the environment, and the article I shared gave me many good examples of this. I do not think it is entirely evil, just growing too fast with too little intention into how adding AI to everything will actually improve them. I would love to hear input on these thoughts and articles as I am also aware that I am quite closed off on this topic, and still want to explore more perspectives.

De-skilling is a big problem, including among highly skilled professionals such as radiologists who come to rely on AI agents to analyze scans. However, if I'm a patient I would prefer the greater accuracy of an AI agent than a human radiologist. Soon self-driving cars will de-skill human drivers, but almost certainly save lives (no drunks, no distracted or tired drivers, no reckless speeders).

It is a human trait to abandon a common skill whenever a technology emerges that provides access to superior results. Writing replaced oral transmission of knowledge, and the ability to remember declined. The introduction of guns lead to the loss of archery and sword skills. Jaquard looms lead to traditional weaving techniques to be lost. Cars undermined the skill in caring for horses and caused the loss of many highly skilled artisnal trades.

AI will cause similar extinctions.

And we humans are relentless in seeking to find ways to increase the amplification of our effort. "Better short-term results for less effort" is built into the 2nd law of thermodynamics and is at the core of evolutionary change - both biological and cultural.

The technical hurdles for useful AI have been passed. The resulting effects can be delayed for a while and maybe directed to some extent, but the changes are inevitable. Even more significant, it seems likely that the barriers to "general" or "super" AI will fall soon, leading to even greater and more unpredictable changes.

A "singularity" is the point past which past experience is an unreliable guide, beyond which we cannot predict. AI generates a major societal singularity.

We know only three things for certain:
1. AI (of some degree) will transform the way human societies organize themselves, as has every major new technology of the past.
2. Both the AI optimists and the AI pessimists are wrong about how the post-singularity world will look (even without general AI)
3. The AI Luddite, like all previous Luddite movements, will fail to stop the singularity.

We will have to be nimble, clear-seeing, and bold if our post-singularity world is to be one that supports human flourishing. I personally am (provisionally) hopeful that we can collectively rise to the challenge.

I think you make some very good points, but I think that this is an oversimplification of the issue. Per my response to Jeanne's comment, AI isn't just an issue of "new vs old technology". I have my four-year degree in Computer Science, and I usually am a fan of innovation myself, however this isn't the whole argument.

AI as it is in the current day takes a lot of energy, a lot of space, and a lot of water, which in a lot of cases is our already limited freshwater. https://weatheronthisday.com/trends/us-drought-2026 https://www.drought.gov/forecasts With a significant portion of the US - some sources say as high as 60% having faced drought this year, the utilization of such high amounts of freshwater for data centers is more than "the price of new technology" and instead a question about the utilization of our limited and sometimes scarce resources.

Furthermore, AI as it is today has very little research, especially long-term, into the impact of data centers on local wildlife and neighborhoods. In my response to Jeanne I shared some articles with people reporting the amount of noise and heat produced by the centers. You can argue that we don't have any studies proving the harm, but we also don't have any studies that prove that data centers are safe. Despite the lack of studies, there are now thousands of data centers across the United States, ranging anywhere from the middle of nowhere Wyoming, to right next to the Nashville Zoo. Furthermore, there have been articles and reports of electricity bills of data center adjacent neighborhoods having electrical bills that are rising or being asked to ration water, bringing us back to the question of resource utilization, and who bears the burden of the centers energy usage.

Another point, again with sources in my response to Jeanne, AI is so new that there has not been time to develop legal precedents. It has been a tale as old as algorithms about "who is accountable when it's generated by a computer?". When AI can make its own decisions, what do we do when those decisions are harmful?

My point is that I need to see these issues resolved before I start touting about how AI is the future, or how "AI is just like every other new technology". Sure, video games and social media have been rotting kids' brains for years, as AI is doing now, but it was never a question about our already dying planet, or who is accountable for doomscrolling (Meta is facing a lot of lawsuits about that right now actually, and they can't blame the chatbot). This isn't just a question of de-skilling, which I agree has been happening long before LLMs were so widespread, but it is about the environment, the usage of freshwater, and the question of "what do we do when things go wrong?". I hope for the best. But I need to see it before I believe it.

I agree completely that we are about to step into the unknown - with any number of risks and rewards, some of which could be quite catastrophic. And I also agree that it is likely that AI will be unlike previous transformative tech (eg. wheel, writing, printing press, radio) in its autonomy, its scope of application, its speed of implementation, and its disruptive power. I disagree that the data centres are a major menace and I suspect that the efforts to halt construction will do next to nothing to slow the development of recursive model self-generated evolution (the ultimate singularity).

Very few actions are ever based on long-term planning (Keynes' "in the long run we're all dead"). Almost all selection processes, both natural and social (and certainly economic) work on short-term differential outcomes. It would take tightly implemented universal restrictions with harsh rapidly-imposed penalties on violations (and unproven suspected violations) to slow the evolution of AI to any significant degree. Without these socially imposed penalties, the short-term potential rewards for companies (and the individuals involved) are just too compelling. Even if some players slow down, others will continue to race towards the dangling rewards.

We simply lack the social, legal, and diplomatic frameworks to materially impede this impulse. There is a vanishingly small chance that the governments of US, EU, China, and others could panic at the cliff-edge we're on and collectively decide to take the draconian action needed to slow things down, but this would be completely illegal in US, EU, and other-law countries. It would also never be seamless. The knowledge exists and there will always be players who will find a way to avoid enforcement. And there is likely more danger in having powerful AI in the hands of only a single group than having competing AIs. This is particularly true if the sole AI is in the hands of those who are sufficiently reckless and anti-social to clandestinely defy the combined will of all major governments.

This analysis suggests that resistance is largely futile, effective only locally and for a very limited time. But this has been the lesson of history. I can think of no luddite movement that has succeeded, even against incentives that were much less powerful and diffuse than we are facing.

So I choose to be provisionally optimistic that in the post-singularity world the benefits will overwhelm the negatives. The alternative attitude is not good for my well-being.