{"id":59232,"date":"2025-04-01T04:36:58","date_gmt":"2025-04-01T01:36:58","guid":{"rendered":"https:\/\/geoconversation.org\/memos\/ethics-in-the-era-of-artificial-intelligence-what-remains-for-humans\/"},"modified":"2025-04-01T04:36:58","modified_gmt":"2025-04-01T01:36:58","slug":"ethics-in-the-era-of-artificial-intelligence-what-remains-for-humans","status":"publish","type":"memos","link":"https:\/\/geoconversation.org\/en\/memos\/ethics-in-the-era-of-artificial-intelligence-what-remains-for-humans\/","title":{"rendered":"Ethics in the era of artificial intelligence: what remains for humans"},"content":{"rendered":"\n

It seems that from every iron they shout: \u201cAI can do this\u201d, \u201cAI can do that\u201d – and now you already feel that you are late. Everyone around seems to have learned to work with neural networks, they are launching their own projects, automate everything that moves<\/a>, and you only have time to wince in irritation. I would like to brush this off, to devalue it: they say, this is all for the lazy and soulless, the person is more important. And to refuse – not to understand, not to delve into, to pretend that nothing is happening. But no, the world is no longer the same. Our life will never be the same. You won’t be able to close your eyes.<\/p>\n\n\n

I’m not going to teach you how to write prom or tell you about the top 10 neural networks. I want to talk about something else – about the principles of working with AI, by understanding which you can easily master this technology. About fears: that AI gives answers based on other people\u2019s, unverified data, or that we will stop thinking for ourselves altogether. About how learning is changing, how we now work with information – and how to preserve ourselves in this new world. Because there are things in which AI will not replace us. Never.<\/p>\n\n\n

You don’t like AI – you just don’t know how to cook it<\/h2>\n\n\n

AI is often presented as a magic pill: open a chat, ask a question, and receive a brilliant answer. But if you go into it with those expectations, disappointment is almost guaranteed. You close the program with the thought: \u201cThis is all nonsense.\u201d I recently developed an editorial policy for our media. And if she just said: \u201cGive me an ed. politics,\u201d the GPT chat would, of course, generate something. But will this text be at least somehow connected with our reality? Hardly.<\/p>\n\n\n

Instead, I started the work this way: I asked the AI \u200b\u200bto help with the structure – to be honest, I myself didn\u2019t fully understand what it should look like. I’m first and foremost geophysicist, not editor<\/a>. And then the real collaboration began. I loaded everything that we already had into the chat: internal rules, messages from work chats, checklists, requirements for texts, examples of technical tasks. Context. Lots of context.<\/p>\n\n\n

And here is the result: what would have taken weeks alone, we did with AI in a couple of weeks – we worked for an hour or two a day. The output is 50 pages of relevant document. Why did I succeed? Because I didn\u2019t ask the AI \u200b\u200bto \u201cjust do it\u201d, but fed him context<\/strong>. And he really started working with me – not according to an average template, but according to my tasks and data.<\/p>\n\n\n

Screenshot of the table of contents of the document \u201cEditorial Policy of GeoConversation. Salt of the Earth\" - includes the main sections and principles of editorial work<\/a>
Editorial policy co-authored with AI: 50 pages, dozens of sections and live editorial context<\/figcaption><\/figure>\n\n\n

And here is the key: you need learn to unload knowledge from yourself<\/strong>. Not necessarily beautiful, not necessarily systematic. Just write it down. Notes, voice notes, screenshots, thoughts on the phone – everything that will help restore the chain of meaning in the future. Creating a personal knowledge base is not a fad, it is the basis of interaction with AI. Because without this, every next request starts from scratch again. And with context, AI becomes a true partner in thinking.<\/p>\n\n\n

AI not only works with context, but also helps see more widely<\/strong>, look at the problem from a different angle. When I was preparing a presentation to apply for a grant for science popularizers, it seemed to me that everything was ready: structure, meaning, visuals – everything was in place. But I decided to show the ChatGPT presentation just in case and asked for a rating.<\/p>\n\n\n

The AI \u200b\u200bhighlighted the strengths, but most importantly, it drew attention to something I hadn\u2019t even thought about: it would be nice to add information about the project\u2019s partners and show site statistics. And it\u2019s true – this immediately strengthened the argument. Without him, I would hardly have seen this gap.<\/p>\n\n\n

In the same way, AI can be used when working on an article – ask it to find contradictions, suggest a metaphor, formulate a point of view \u201con behalf of\u201d a hypothetical reader. All this helps add nuances to the material<\/strong>, look at it from different angles. Of course, the final decision and verification remain with the person, but The potential for collaboration is huge<\/strong>.<\/p>\n\n\n

People are afraid of new things. And that’s okay<\/h2>\n\n\n

New things have always frightened humanity – and AI is no exception. But before we panic, let’s honestly ask ourselves: is it really that scary, or do we just not understand how to handle it? Fear is normal, especially in the beginning. The main thing is not to get stuck in it.<\/p>\n\n\n

The ruler, ink and calculator were already there – now it\u2019s AI\u2019s turn<\/h2>\n\n\n

There is a fear that a person will begin to rely on neural networks and become completely stupid. The thought is frightening: \u201cwhy think if you can just ask?\u201d But this is not the first technological leap in history. Previously, people counted in columns, then calculators appeared. Geologists and geophysicists spent months drawing maps with ink and drawing isolines by hand – and now a soulless machine does it for us in five minutes. Have we really gone dumb? No. Just stop wasting time on routine<\/strong> and began to work on things where the brain is needed, and not a ruler and whatman paper.<\/p>\n\n\n

It will be the same with AI. He will take away the mechanical part of the work from us – and that\u2019s normal. But the excuse \u201cI don\u2019t have time to think\u201d won\u2019t work anymore. Vice versa, thoughts, strategic thinking, criticism, analysis – that\u2019s what will become more important<\/strong>. AI can process data, sort, format – but conscious understanding of tasks, the ability to formulate goals and see consequences – this is still a person\u2019s area of \u200b\u200bresponsibility<\/strong>.<\/p>\n\n\n

This is the main risk – not the AI \u200b\u200bitself, but passive attitude towards him<\/strong>. If you outsource everything to a machine, your brain stops training. If we completely shift the work to AI and stop using our heads, then we lose the role of an active participant in the process – we become passive users who simply watch what is happening. But if we use AI consciously, as a tool, and not as a replacement for ourselves, it can not weaken, but on the contrary, strengthen our abilities, make us smarter, and make our work deeper.<\/p>\n\n\n

\"A<\/a>
Once upon a time, maps were hand-drawn with ink – now AI does this. But the important conclusions still remain with the individual.<\/figcaption><\/figure>\n\n\n

The pendulum will swing back: why human experience will become even more important<\/h2>\n\n\n

One of the most interesting – and perhaps most underestimated – worries: AI will start educate yourself<\/strong>. Already, generative models are largely based on data created by other generative models. Synthetics devour synthetics, and the further it goes, the more the AI \u200b\u200bloses touch with reality. It’s like learning life from books without going outside.<\/p>\n\n\n

But this is not a reason to panic – on the contrary. Human experience, real data, personal notes, cases and experiments – all this will be valued many times more.<\/strong> Because only a person can bring something new, real, non-fictional into the world. This means that our task now is not just to \u201ckeep up with AI\u201d, but to accumulate and comprehend our own experience.<\/p>\n\n\n

This is the answer to fear: develop yourself<\/a>, invest in knowledge, act.<\/strong> Learning through mistakes<\/a>, through real practice. You can\u2019t understand the world without touching it. And yes, don\u2019t forget about the base. Everything that seems like a boring school routine – the multiplication table, Newton’s laws, basic skills in working with the world – actually shapes structure of thinking<\/strong>, onto which our experience is then superimposed. Without this foundation, it is impossible to think deeply, critically, or\u2014especially important in the age of AI\u2014understand when a machine is wrong.<\/p>\n\n\n

AI can generate anything, but only man can live<\/strong>. And the more synthetics around, the more valuable becomes what is truly yours.<\/p>\n\n\n

\"Ouroboros<\/a>
AI is eating itself. When neural networks are trained on synthetic data, the value of human experience only increases.<\/figcaption><\/figure>\n\n\n

I don\u2019t trust AI – and I\u2019m doing it right<\/h2>\n\n\n

I myself don\u2019t always trust AI. He can quickly collect information, highlight useful thoughts, but guarding quality – always a person<\/strong>. I think of myself at such moments as an internal quality control department. Everything the AI \u200b\u200boutputs goes through me.<\/p>\n\n\n

Here are a couple of simple rules that will help don’t miss a mistake<\/strong>:\u2013 Facts<\/strong> (dates, names, numbers) – I always check. If there are links, I go to the source.\u2013 Opinions and interpretations<\/strong> \u2014 I perceive it as another point of view. It\u2019s like I\u2019m consulting with a smart colleague who isn\u2019t always right.\u2013 Recommendations<\/strong> (for example, \u201chere\u2019s your daily plan\u201d) – I evaluate it in practice. Sometimes it fits, sometimes it doesn’t.<\/p>\n\n\n

And one more life hack – always ask for clarification<\/strong>, why AI offers this particular solution. Firstly, he himself \u201cthinks\u201d better when he explains. Secondly, you begin to understand logic and test it. It also improves the skill of asking questions, and this, by the way, is a key competency of the 21st century.<\/p>\n\n\n

AI doesn\u2019t just help you work\u2014 it teaches reflection<\/strong>. When you don\u2019t just read a text or watch a video, but ask yourself questions:<\/p>\n\n\n