Great context, Maryan. Technical support is a good example of the one thing AI can do because it is the child of geeks and software businesses. There's 90% of what matters in writing that AI can not do because it has been in nobody's economic interest to spend the time loading in the data to train it on those 90% other things. Follow the money. You gave a good example of "nothing lost, nothing gained" when it comes to AI. Ask AI to perform low-risk chores that nobody will see other than you. If anything good happens--hooray. If AI is a big fizzle, move on to other methods for getting work done.
I am very impressed by the information you are providing. I recognize that I am part of a group of innocent individuals who occasionally seek assistance from AI, but I do not believe I have consciously committed any ethical violations.
AI and its evolution -- a complicated and interesting phenomenon. Each of us has to consider how and why we use it. And we have to keep in mind that once upon a time, electricity was widely thought to be an instrument of the devil. We just keep on keeping on, I guess. As always, thanks for being here, Maria.
Thank you Maryan for the thoughtful commentary. Trade offs. Always trade offs.
My default mentality is slow to support more regulation. In this complex case however it sure seems like we need some safeguards.
My selfish concern is AI generated content being passed off as authentic, old school real writing by real people. What if, like with food ingredients, there were labeling requirements for creative work? Some sort of disclosure requirement that made it clear if (maybe even how much) AI content was included.
Really good ideas, Craig. Actually, the US Copyright Office (European, too) mandates that all IPs must have a disclosure as to the amount of AI input. And, they will not register rights on any property that is more AI than human. The problem is, no one using AI is going to or has been known to disclose such use. It's the thought that counts, I guess. I am 100% onboard with your "selfish concern." So glad you're here.
I, too, have learned to see AI as a valuable option for IT support! AI has walked me through a number of highly technical issues as I navigate the worlds of Medium and Substack, as well as integration with Wordpress. Great article; full of common sense.
Interesting viewpoint but a conventionally limited one. So AI ingesting writers, artists and creatives is unethical but then what about the absolute droves of others. The AI steps to help you with your technical problem came from ingesting the work of technologists, technical writers and small businesses. Is this not unethical too? This seems to be one of many inconsistent attitudes that we must better grapple with in dealing with AI. However, I’m a pessimist and we won’t.
I see several others on Substack who love AI, proclaim its virtues, seem depressed by what tasks it will take of their writing jobs. Their perceptions and experiences do not jibe with mine at all. I did some experiments with deeper research on AI with a great AI journalists. It had to be in a field I was an expert in. The report was long. It was complete. But it was like a flat encyclopaedia of everything on the subject. No priorities, nothing very human about it, et al. It is one of the reasons I started writing Techtonic: to point out the ongoing gross errors in all of technology, but especially AI
The trend I notice is that people, like myself, with decades of experience with technology (except for techbro fanboys) are much more skeptical and measured. It is the more innocent who are being metaphorically slaughtered
Certainly it is all IP being scavenged and consumed by AI, not just creative works. But for things I've done and how I understand Maryann's use of AI to problem solve, I believe I'm asking AI to scan available public resources, websites, and give me an ordered, useful summary of what I need. I, personally, would not knowingly use AI to infringe on anyone's content rights. I think many predators are, however.
What do you mean about the Innocents being slaughtered? I'm curious.
3. Is it - through the ingestion of IP - directly affecting someone’s income stream
The innocents remark is about the naïveté that I see exhibited by people without a deep IT design, build, implement background. So for instance, you think all the AI is doing is scanning publicly available databases but you / no one can be sure of that. Depending on the particular model it may include “references” but theses are added ex post so many not be how its probability engine works. Another example directly related to MaryAnn’s usage that I have found that it gives you technical steps / videos from some person’s monetized Youtube channel. So it has taken their IP with a monetary Impact as well. The problem of mis/disinformation is profound and what is more important uncertain. AI is unpredictably unpredictable.
I can't argue that. Is it kind of difficult for you to know who has a deep IT background and who does not? And I always feel concern when any of us make generalized statements about what people think, or know, or feel--but that's just me. I'm very interested in your opinions here.
Yes, I think I see what you mean. I have a lot of free content online that is there in order to help those I can help discover me, weigh my helpfulness to them, and hopefully begin a relationship between content provider and content seeker. AI removes any chance of that relationship.
I have been primarily uncomfortable with all of it, but the handful of times I've used it, I have found it to be such a time saver that I know I will use it again. But your remark does increase my wariness.
I think the future of human-AI relationships will be really interesting to observe. Our tendency to anthropomorphize and make friends of our technology is in turns amusing and disconcerting.
I use AI but I am very skeptical and thorough reviewing the results. Most of the innocents I mentioned (my newsletter goes into many more examples) don’t have any of that. It looks like magic to them. There are many robust scientific studies that show how quickly people lose their critical skills, treat AI output as remarkably accurate, and other poor behaviors. The latest studies of ChatGPT 4.5 show that 1/3 of its output is erroneous. Here is a quote from Gary Marcus: “If however, you simply "trust" the Al outputs due to lack of knowledge, skill, or willingness to review results, the long term damage will outweigh the initial productivity gains you got so "hyped" about.” That is my slaughter of the innocents
I use AI to replace Google search--which has just become a complete waste of my time. I vacillate between GPT, Gemini, and less often--Copilot. Interestingly, they are all changing so quickly that I'll find GPT to be more accurate and faster one day and Gemini the next. Copilot is the least accurate overall. Gemini's newest release has some interesting skills. And don't forget, Amazon (the king of beasts) is about to release Alexa+ which they claim will run rings around all others. Uh huh. I'm quite sure.
Yes Google Search has been bad for maybe 7 or 8 years. I switched to Quant but it was hit or miss. AI search is reasonable for many simple things. CoPilot is not very good. I was a Microsoft implementation partner for 20 years. It is another example of where their marketing is many years ahead of the software product
I expect I've always been conventionally limited. Instruction manuals, while often copyrighted, are widely disseminated by manufacturers and stakeholders as a marketing technique. So collecting that factual, not very creative info feels different to me--but that's just me. Yes, it's a conundrum. No, none of us has all the answers. My background is strongly tech, too, and I love burgeoning technology in general. Thanks for your input.
I'm with you. I have found AI incredibly useful for tasks that would take me hours, like scouring the web to gather information or some kinds of copywriting for media. But I would never use it for my creative work or anything I consider my IP. I also would never upload my IP (such as a novel or story) for it to help me (as some writers do for generating cover copy or market research), even knowing anything I've published digitally has probably been used without my knowledge or consent. It is astounding how quickly everything is pirated. I hope more people will have these conversations and we'll see AI develop as a useful tool for us all without an ongoing threat to human IP.
Interestingly, I read that Microsoft is closing many if not all of it's AI datacenters. Some experts have said it feels like MS is stepping back from the initial all-out embracing of AI, but who knows why that company does anything. On the other hand, other key players are investing more. Wait and see is all we can do.
Indeed. Wait and see. I have wondered if it could ever come to artists (and others) eschewing digital mediums, e.g. only selling print books, to avoid the cannibalisation of works.
Great context, Maryan. Technical support is a good example of the one thing AI can do because it is the child of geeks and software businesses. There's 90% of what matters in writing that AI can not do because it has been in nobody's economic interest to spend the time loading in the data to train it on those 90% other things. Follow the money. You gave a good example of "nothing lost, nothing gained" when it comes to AI. Ask AI to perform low-risk chores that nobody will see other than you. If anything good happens--hooray. If AI is a big fizzle, move on to other methods for getting work done.
That about says it.
I am very impressed by the information you are providing. I recognize that I am part of a group of innocent individuals who occasionally seek assistance from AI, but I do not believe I have consciously committed any ethical violations.
Thank you for your interesting contributions.
AI and its evolution -- a complicated and interesting phenomenon. Each of us has to consider how and why we use it. And we have to keep in mind that once upon a time, electricity was widely thought to be an instrument of the devil. We just keep on keeping on, I guess. As always, thanks for being here, Maria.
Thank you Maryan for the thoughtful commentary. Trade offs. Always trade offs.
My default mentality is slow to support more regulation. In this complex case however it sure seems like we need some safeguards.
My selfish concern is AI generated content being passed off as authentic, old school real writing by real people. What if, like with food ingredients, there were labeling requirements for creative work? Some sort of disclosure requirement that made it clear if (maybe even how much) AI content was included.
Really good ideas, Craig. Actually, the US Copyright Office (European, too) mandates that all IPs must have a disclosure as to the amount of AI input. And, they will not register rights on any property that is more AI than human. The problem is, no one using AI is going to or has been known to disclose such use. It's the thought that counts, I guess. I am 100% onboard with your "selfish concern." So glad you're here.
Thank you for scoop on existing law. I didn’t know. Let’s stir the pot pushing for enforcement.
I'd love to. But first we have to make all humans truthful...
I, too, have learned to see AI as a valuable option for IT support! AI has walked me through a number of highly technical issues as I navigate the worlds of Medium and Substack, as well as integration with Wordpress. Great article; full of common sense.
Interesting viewpoint but a conventionally limited one. So AI ingesting writers, artists and creatives is unethical but then what about the absolute droves of others. The AI steps to help you with your technical problem came from ingesting the work of technologists, technical writers and small businesses. Is this not unethical too? This seems to be one of many inconsistent attitudes that we must better grapple with in dealing with AI. However, I’m a pessimist and we won’t.
I see several others on Substack who love AI, proclaim its virtues, seem depressed by what tasks it will take of their writing jobs. Their perceptions and experiences do not jibe with mine at all. I did some experiments with deeper research on AI with a great AI journalists. It had to be in a field I was an expert in. The report was long. It was complete. But it was like a flat encyclopaedia of everything on the subject. No priorities, nothing very human about it, et al. It is one of the reasons I started writing Techtonic: to point out the ongoing gross errors in all of technology, but especially AI
The trend I notice is that people, like myself, with decades of experience with technology (except for techbro fanboys) are much more skeptical and measured. It is the more innocent who are being metaphorically slaughtered
Certainly it is all IP being scavenged and consumed by AI, not just creative works. But for things I've done and how I understand Maryann's use of AI to problem solve, I believe I'm asking AI to scan available public resources, websites, and give me an ordered, useful summary of what I need. I, personally, would not knowingly use AI to infringe on anyone's content rights. I think many predators are, however.
What do you mean about the Innocents being slaughtered? I'm curious.
There are several issues:
1. What do we mean by creative?
2. It is all IP but is it public or private
3. Is it - through the ingestion of IP - directly affecting someone’s income stream
The innocents remark is about the naïveté that I see exhibited by people without a deep IT design, build, implement background. So for instance, you think all the AI is doing is scanning publicly available databases but you / no one can be sure of that. Depending on the particular model it may include “references” but theses are added ex post so many not be how its probability engine works. Another example directly related to MaryAnn’s usage that I have found that it gives you technical steps / videos from some person’s monetized Youtube channel. So it has taken their IP with a monetary Impact as well. The problem of mis/disinformation is profound and what is more important uncertain. AI is unpredictably unpredictable.
I can't argue that. Is it kind of difficult for you to know who has a deep IT background and who does not? And I always feel concern when any of us make generalized statements about what people think, or know, or feel--but that's just me. I'm very interested in your opinions here.
Yes, I think I see what you mean. I have a lot of free content online that is there in order to help those I can help discover me, weigh my helpfulness to them, and hopefully begin a relationship between content provider and content seeker. AI removes any chance of that relationship.
I have been primarily uncomfortable with all of it, but the handful of times I've used it, I have found it to be such a time saver that I know I will use it again. But your remark does increase my wariness.
I think the future of human-AI relationships will be really interesting to observe. Our tendency to anthropomorphize and make friends of our technology is in turns amusing and disconcerting.
I use AI but I am very skeptical and thorough reviewing the results. Most of the innocents I mentioned (my newsletter goes into many more examples) don’t have any of that. It looks like magic to them. There are many robust scientific studies that show how quickly people lose their critical skills, treat AI output as remarkably accurate, and other poor behaviors. The latest studies of ChatGPT 4.5 show that 1/3 of its output is erroneous. Here is a quote from Gary Marcus: “If however, you simply "trust" the Al outputs due to lack of knowledge, skill, or willingness to review results, the long term damage will outweigh the initial productivity gains you got so "hyped" about.” That is my slaughter of the innocents
I use AI to replace Google search--which has just become a complete waste of my time. I vacillate between GPT, Gemini, and less often--Copilot. Interestingly, they are all changing so quickly that I'll find GPT to be more accurate and faster one day and Gemini the next. Copilot is the least accurate overall. Gemini's newest release has some interesting skills. And don't forget, Amazon (the king of beasts) is about to release Alexa+ which they claim will run rings around all others. Uh huh. I'm quite sure.
Yes Google Search has been bad for maybe 7 or 8 years. I switched to Quant but it was hit or miss. AI search is reasonable for many simple things. CoPilot is not very good. I was a Microsoft implementation partner for 20 years. It is another example of where their marketing is many years ahead of the software product
I expect I've always been conventionally limited. Instruction manuals, while often copyrighted, are widely disseminated by manufacturers and stakeholders as a marketing technique. So collecting that factual, not very creative info feels different to me--but that's just me. Yes, it's a conundrum. No, none of us has all the answers. My background is strongly tech, too, and I love burgeoning technology in general. Thanks for your input.
I'm with you. I have found AI incredibly useful for tasks that would take me hours, like scouring the web to gather information or some kinds of copywriting for media. But I would never use it for my creative work or anything I consider my IP. I also would never upload my IP (such as a novel or story) for it to help me (as some writers do for generating cover copy or market research), even knowing anything I've published digitally has probably been used without my knowledge or consent. It is astounding how quickly everything is pirated. I hope more people will have these conversations and we'll see AI develop as a useful tool for us all without an ongoing threat to human IP.
Interestingly, I read that Microsoft is closing many if not all of it's AI datacenters. Some experts have said it feels like MS is stepping back from the initial all-out embracing of AI, but who knows why that company does anything. On the other hand, other key players are investing more. Wait and see is all we can do.
Indeed. Wait and see. I have wondered if it could ever come to artists (and others) eschewing digital mediums, e.g. only selling print books, to avoid the cannibalisation of works.