@reformed_content_marketer@mastodon.social

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@knowprose@mastodon.social @mcc@mastodon.social I don't think it's truly ethical to use someone's data for LLM training just because "you technically agreed to this in the terms of service in 2014" or "you contributed to a creative commons project in 2009" or "you were an employee in 2018 and therefore we can do whatever we want with your words even though you never agreed to contribute to our AI and that wasn't in your contract." Almost no one could have understood what they were supposedly consenting to back then.

4

2

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@knowprose@mastodon.social @mcc@mastodon.social what are they? what are the ethical standards? I've seen a few LLMs that were better than most, but none that were quite what I'd consider "ethical." I'd like to see one based on true consent, i.e. direct opt-in data collection only from people who know what they're contributing to.

4

3

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@knowprose@mastodon.social @mcc@mastodon.social I feel weird about even public domain. Again, no one could have guessed what they were "consenting" to before recent years. I shared photos under (I think?) a CC license a long time ago, when that just meant anyone could post them w/o attribution and maybe photoshop them with a lot of work. I wouldn't have if I'd guessed they'd be used to train super-fast deepfake generators. I do think public domain dataset LLMs are far better than most, though. Less unethical.

3

2

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social AI has displaced some jobs, but more often it's turned jobs (or parts of them) into checking LLM output. That's actually worse than just getting rid of the job in some cases, because it leads to producing more misinformation. Most LLM hallucinations are subtle, so you have to closely check every bit information to catch them. Very people are as careful as they think they are. Fact-checking is often mind-numbing work, so it's hard to stay vigilant. 🧵 (2/ ?)

2

1

2

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social for one, because the tractor didn't screw up the information ecosystem on top of displacing laborers. We now have LLM hallucinations in scientific research journals, government intelligence that affects military decisions, legal documents, product manuals and informational materials for things that might be dangerous if handled poorly, educational resources, and healthcare systems. 🧵 (1/ ?)

2

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@celesteh@hachyderm.io @farseen@fedi.nptr.xyz @knowprose@mastodon.social @mcc@mastodon.social yeah, I wish everywhere had the concept of moral rights for creators. Copyright violation doesn’t encompass most of the harms being done. AI training isn’t just (IMO not even primarily) a property rights issue. We need a larger framework that takes into account moral injury issues, dignity in labor, etc.

2

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social I think there are a lot of other things adding to that gut-wrenchingly wrong feeling. You probably care how your work impacts people, and you see all these little ways LLMs have made it worse for end users or other workers. Maybe you’ve had to trade fulfilling work for depressing work where it’s harder to make something you can feel proud of. Maybe there are more barriers to human connection, and more things I can’t articulate either.

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social Even if you try to be careful, your boss probably thinks fact-checking should take less time than it actually does. And many people don't really check things because they're overwhelmed with too many tasks, don't read well, or don't think the errors will matter much. And many organizations have calculated that it will be more profitable for them to sacrifice a little information accuracy for "efficiency." 🧵 (3/ ?)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social LLMs are also cutting into journalism, as it it weren't already too underfunded. That affects our ability to hold power accountable and have a functioning democracy. Even before LLMs, the internet had already kind of screwed up the funding model for journalism. The past few decades have seen tons of job cuts, newspaper closures, and the remaining journalists doing more work for less pay, often with bosses forcing clickbait for financial reasons. 🧵 (4/?)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social When people get news from an AI overview or a chatbot, the news sources AI draws from get fewer clicks and less ad revenue. So now we're entering a world with even fewer people doing important investigative work well. We can expect more corruption and even less shared understanding of reality. 🧵 (5/?)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social a small, close-to-home example of the LLM info problem: my relative called her doctor's office to ask which vaccines they recommended for her this year and whether she needed to schedule an appointment to get them. She couldn't get through to a human on the phone, and the AI didn't answer the question. So she then messaged the doctor's office. And then... 🧵 (6/ ?)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social They made her wade through another LLM with a disclaimer about how its info may be inaccurate before she could finally message her "care team." After that, a "nurse" finally replied... with a clearly AI-generated answer that assumed she might be living across the country (never mind where the doctor's office is) and still didn't answer the question. After that, she gave up and decided to wait for her annual physical to ask her doctor in person. 🧵 (7/ ?)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social As a result of the doctor's office's LLM usage, she'll be getting whichever vaccines she needs later than she otherwise would. If this kind of thing is becoming the norm, more people will die as a result of people waiting longer to get vaccines or not getting them at all, just because the doctor's office group decided to be "efficient" by using LLMs to answer patient questions. 🧵 (8/9)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@farseen@fedi.nptr.xyz @celesteh@hachyderm.io @knowprose@mastodon.social @mcc@mastodon.social If LLMs were just displacing some knowledge workers, that would be one thing. Having a writing-related job was a privilege. What I can't stand is seeing that work replaced with bad information. I care deeply about these issues. Knowing that the machines used to create or amplify them were trained on my own work, which I always tried to do with care for information accuracy, is just an extra punch in the gut. I don't want to contribute to this shit. 🧵 (9/9)

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1w ago

Replying to

@knowprose@mastodon.social @mcc@mastodon.social public domain is a legal framework. We're talking ethics. Legal does not equal ethical, and some things in the public domain were produced by living people who waived copyright. My point is that consent requires understanding. You can't fully consent to something that doesn't yet exist. If you dedicated something to the public domain in 2002 or worked for the gov, and the work was later used for chatbot training, you didn't truly consent to that. It's still an ethical gray area.

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 2mo ago

Replying to

@Em0nM4stodon@infosec.exchange There have been plenty of times when an AI "AI-detector" claimed my completely human-written content was partially AI-generated. These tools are so stupid, yet so many people trust them. I wish there were a reliable, non-privacy-invasive way of assuring human provenance so we could stop these kinds of false accusations. I want there to be reliably AI-free spaces again. All I can really do is use the trust system and hope that showing someone a doc's history occasionally is enough

8

2

3

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 4mo ago

Replying to

@hbons@mastodon.social @thomasfuchs@hachyderm.io It's a step forward, but it's not good enough. When I searched for information about that law, one of the first results was a marketing firm called Weventure talking about how they can "make sure your AI-powered content doesn't have to come with a warning label." I'm not linking to them because I don't want to boost their SEO. Basically, they're going to do just enough human review to get around the law.

9

3

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 4mo ago

Replying to

@hbons@mastodon.social @thomasfuchs@hachyderm.io If my past experiences in marketing are any indicator, people aren't going to check very closely for AI hallucinations before publishing articles that won't be marked as AI. Taking the time to do proper fact checking is obnoxious and not "efficient" enough.

3

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 2mo ago

Replying to

@bms@mastodon.bsd.cafe @mcc@mastodon.social @lcamtuf@infosec.exchange am I seeing right that anyone globally can opt out, not just UK publishers? And are other tech companies following suit? This is giving me hope that maybe I'll once again be able to have personal website projects that aren't used for AI training or automatically, poorly "summarized" by AI. really hoping!

1

0

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 4mo ago

Replying to

@hbons@mastodon.social @thomasfuchs@hachyderm.io still happy to see that the EU is trying to do something about unmarked AI content, though! Hopefully non-EU countries will follow suit, and people will keeping working on ways to deal with all the AI slop/hallucinations floating around the internet.

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1mo ago

is it another form of ai psychosis if you're going crazy due to everyone else's obsession with ai?

(but also, i didn't get enough sleep last night, and maybe that's all this is)

(who am i kidding, i've been going crazy way too long for this to be resolvable with a good night's sleep)

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1mo ago

i miss when i was enthusiastic about life

i miss thinking i had a bright future

i miss thinking people valued my skills

i miss not knowing how llms were trained

but more than that, i miss thinking that most people cared at least a little bit about information accuracy, and that our democracy was strong, and that things were generally going to get better

maybe they still will, but it seems a lot less likely these days

i think we're in for a rough rest of the decade at least

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1mo ago

When I sound angry about AI, I'm not really blaming the average AI user.

I'm not blaming people who feel forced to use generative AI at work, especially when it's the new industry standard and your boss mandates it. (I've been there, too). I'm not blaming people who think they need to use generative AI to avoid being left behind. (I've been caught in that trap, too). I'm not blaming people who are just curious about generative AI and don't really know how it was built. (I was once there, too).

0

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 1mo ago

What is up with the whole "can't put the genie back in the bottle" thing? Who thought that was a good metaphor for anything?

0

1

0

why did i ever do content marketing @reformed_content_marketer@mastodon.social

· 2mo ago

Replying to

@n_dimension@infosec.exchange @BenjaminHCCarr@hachyderm.io i read something a while back about how a company was telling job applicants to do exactly that exercise during job interviews. If you wanted the job, you had to open your chatbot account and prompt it to assess you, then show that to the interviewer. Presumably you didn't get the job if you didn't want to do that or didn't have a chatbot account. Scary AF

0

1

0