#Using AI to help me lol

[elliot-the-frog]

I have been doing my history homework for the last 4 hours. Due tomorrow

[selfportrait27]

HD version of a classic Ween photo (probably AI upscaled). 1992-ish.

[adulthumanproblem]

AI can't replace humans in language learning or even translation

Google translate is already only really useful in a pinch

Because here's the thing: you need to under culture as well

Take Japanese for example:

It has multiple levels of politeness. You can learn all the words and grammar just fine, sure, but without understanding why, when and for whom you are using which? Doesn't really get you anywhere

We all know there are many ways to say "I" in Japanese. That's largely dependent on who you are talking to and what your position in relation to theirs is

I thought "watakushi" was just a different way of saying "watashi", maybe depending on the region (like "atashi", which is mainly used in Tokyo and female coded)

But it turns out, "watakushi" is polite language, something you'd use towards a client or customer

There are different words for things like directions in polite language as well

And while it sounds complicated, I'd argue most languages do this. You speak differently with a client than you would a friend

But it's still something you need to know. While your friends may not care as much, a client or potential employer can absolutely be put off by it

That's why it doesn't really help to just repeat words and phrases

Culture and language are intertwined

English is used differently by every country, every region where it's spoken as well

[thelaurenshippen]

open ai made a little post about how they chose their chatbot voices and two things stood out to me:

a) I am genuinely pleased and surprised that the voices come from specific real people that they hired and paid, rather than being built off of all of the data open ai has scraped. the voice for siri did one job for one company 20 years ago and now she is literally everywhere without ever being compensated by apple or even acknowledged as the voice. there's simply no way she was paid enough in that original job. whereas, open ai says "each actor receives compensation above top-of-market rates, and this will continue for as long as their voices are used in our products." this could mean literally anything (what market rate are we going above? does continue mean they get residuals or get paid for doing more sessions?) but, christ, at least the actors are doing it with full knowledge of what their voices are being used for and can decide for themselves if the compensation is enough

b) the post talks about working with "award-winning casting directors" to get the voices. first of all, yay for paying casting directors! we love to see it. but they also say the CDs received over 400 submissions in a week and they state that like a big number and it's just...not at all. and look, without knowing the intricacies of the casting process, it's hard to know what approach the CDs took - it's very possible they were selective from the jump and 400 is a lot from the pool they were tapping. 400 would be a lot if you were going to the agencies directly and asking for names, but there's just no way in hell open ai went after big stars for this. so it would've been a pool of unknowns. in which case, 400 is laughably small. even if you're not using the big casting sites like actorsaccess, I've worked on projects with CDs and their own internal systems where we've gotten over 100 submissions for a single role. I've posted roles on casting sites and received literally thousands of submissions in just a few days

look, I have a very limited perspective on this - I am not a casting director (imo, one of the most important and undervalued jobs in hollywood) and I, in fact, hate the process of casting with a passion. but 400 just seemed like such a tiny pool to pull from and, idk, it heartens me! it's heartening to think that there's very little interest from actors and agents to be doing this kind of stuff. and absolutely no shade to the actors who did--I want actors to get their bag however they want as long as it's, like, safe sane and consensual, you know? but there's something encouraging about thinking that open ai hired some big casting agency to get their foot into the voice acting door and people didn't come running

[bratbarzal]

Hey, sorry for bringing this up but I was wondering if you could give any advice on how to spot AI in fic or writing? It makes me feel weird to think I'm reading or enjoying fics on here and the content is AI generated :(

I'm not necessarily the one to be policing this but I'll do my best to help bc I understand how disheartening it is to think you're unintentionally consuming AI generated content when it doesn't align with your own morals. I'll do my best to be helpful, but if anyone has any other advice, please feel free to add on if you can!

I'll also preface these points by saying, I'm talking about using AI to write a fic. I'm not talking about AI to translate, I'm not talking about AI to generate a prompt, or any other ~innocent use of AI within this sort of space. I don't intentionally use it at all as a personal decision/dislike, but I can understand it's place and the fact that it can be unavoidable for certain things that there is no ~easy alternative to.

Imperfection! I don't want to sit here and say perfection in writing is anything to be alarmed by - people proof read their work, or have people who do so for them, to make sure it flows or reads as well as it can, and obviously no one wants to be putting out anything that isn't their very best - but if I see any sort of mistake in writing, I feel safe enough to assume it wasn't generated by AI. Spelling errors, grammar mistakes, this boy doesn't even know the difference between there, their, and they are sort of thing. Imperfections are human. Ramblings are human - does a sentence go on way too long (me)? Does it seem like the writer is running out of ways to avoid overuse of the same word and are getting a migraine from scrolling through power thesaurus (me)? All indicators of someone who doesn't know when to stop, and isn't interested in keeping things looking perfect, therefore most likely not AI. The same can also be said about imperfections in written speech. Stutters, hums (uhms, ah's, ehh's, uhh's etc), saying phrases that are incorrect (like "could care less" which should always be "couldn't care less" lmao) are all human, and would not be replicated in speech generated by AI.

Personality! I think it's impossible to write something without inserting a little bit of yourself into it, and often you can see a consistency between someone's personal posts, and their writing. I acknowledge I'm saying this as someone who signs every sentence off with lmao like it's a period, and then writes 81k word fics with excessive use of the oxford comma and that often read like unhinged ramblings, but what I'm saying is, you will see their personality in their writing. I'm admittedly and to a fault sarcastic, and a raging idiot, and the shit I say on here reflects that, but my writing, and the people I write specifically, reflect that, also, and do so in a way that shows there is a human being behind it. I've never read an AI text that isn't robotic, or monotone, or lifeless. It doesn't matter how many prompts you give it, or how many times you run it through the mill, it cannot manipulate the essence of a person. I also recognise that not everybody shitposts in the way I personally do, but if you can get a good read on a person through their blog, through posts they make otherwise, or tags they add, or asks they answer, you should be able to tell if their writing is their own.

Language! Similar to the above, is the language used throughout consistent with the kind of language that writer uses elsewhere. Is any of it unnecessarily formal? Does the fiction that you are reading, read like fiction, does it read like there is thought and intention behind every word as there would be behind something a real person wrote?

Emotion! Again, tying into the above few points, is there more to the story than just action? I see a lot of AI that is - this happens, this happens, this happens, speech, movement, telling and rarely showing. It's all very specific, and again, robotic. Shoutout katie @nol-pat bc we were just discussing examples and she brought up a valid point which is that there is never anything beyond the action of it all, because the bot isn't capable of the emotion or explanation that goes along with any of it. There is little thought, there is little feeling, there is little grit, if that makes sense? And sometimes, people write this way, but if they have the power to be writing a story, they usually have the power to add beyond the bare minimum of basic human movement. And in a similar vein, does what you are reading make YOU feel something? Beyond a description of flushed cheeks and heart hammering, can you really relate to what you are reading, to how the characters are reacting and responding to what is happening? Do you feel immersed in what you are reading, and do you feel that way throughout?

Consistency! Chances are if you are reading anything multi-chapter or that is a continuation of a previous work, it won't be AI generated. If there is a story you can follow through each chapter, if there are call-backs and references and repeated traits and characteristics, that make the story flow, there is someone putting a whole lot of pain-staking thought behind it. I think the same goes for generally longer fics. From what I've seen of AI in other works, the longer it is, the more likely it is to go off the beaten path. There will be inconsistencies later down the line in an AI fic because of the way it is set up - it usually takes more and more prompts to make it longer, and the more you instruct it to expand, the more it forgets what it originally gave you.

This last one I was hesitant to bring up because I see it all over, and I see people pointing the finger at it's use, and I don't generally agree that the EM-dash is the devil, but in looking at a few proven and assumed AI works while I've been thinking about this, I feel like it's something I want to raise for you to look into yourself, because it is very much something that keeps cropping up.

I'll preface what I'm about to say with I don't think the use of an em-dash necessarily an outright indicator, but bare with me on this whole thing, because there is something to say about its overuse.

-, a dash, is typed outright, and a part of most regular keyboards, certain writing tools will autocorrect this to a longer version, which is an EN-dash (e.g. I use pages, when I type a -, it will often elongate it to an – within the app. - vs –) when it is used to signify an "up to" or a "until" like you'd see on opening hours or something, but when works are copied into Tumblr, it will usually just revert to it's regular dash size. An EM-dash is a longer dash and the correct use of this would be to indicate a break/change in sentence structure —. So that is - vs —.

- which is a dash/hyphen (eg. book-end, co-worker, etc.)

– which is an EN-dash (eg. 9am – 5pm, pages 4 – 5, etc.)

— which is an EM-dash (I'll use an example from merriam-webster here, because as I will go onto say, I would usually just use a regular dash “That is—I suppose it is concerning.”)

If you type in a DOUBLE hyphen into most tools, it will correct to an EM-dash, and I don't know if it's the tool I'm using, but when I have then copied this into Tumblr, or written straight into Tumblr if you're the kind of live-life-on-the-edge maniac insane enough to rely on doing that, it usually ends up coming out as a double hyphen (--, so that's -- vs —). Again, I'm not saying em-dashes are a certain indicator, but something to consider if the rest of the text seems off. There's a lot of research and debate around this, but in my personal experience, as a creative writer, as a reader, as a general enjoyer of fiction, when I see an EM-dash outside of a corporate email or an essay, it feels generally misplaced, and you'll notice it replacing commas, colons, semi-colons, etc and you will be able to see what I mean for yourself. I (again, PERSONALLY) have always just written a hyphen, never even a double that would auto-correct, which is why it's shifty to ME. An em-dash, despite it's definition and it's place in grammar and punctuation as a whole, reads more like a personal or professional function, and not something I would see or use in creative writing. I'll attach some reading on this, I've tried to find anything not behind a paywall, and a few different opinions on the matter. (1) (2) (3) (4)

Something I will say to look out for specifically when it comes to the EM-dash, is how it is being used. If you think it's being used to cover up where a human imperfection may be, especially in speech, like a stutter, or a pause to think, or fillers like "uhh" or "erm" you have your answer.

Overall, AI is a complex and ever-evolving beast, and it's borderline impossible at the end of the day to ever know for sure, but I can tell you as someone who has been on this site a long time, there are a lot more creatives on here than there are those who are fraudulently posting AI works as their own. Sometimes it can feel like a plague, and sometimes the tags are riddled, but if you're ever unsure, ask around. People will direct you to fics that made them feel something - fics that made them laugh or cry or think, and you'll find the authentic writers behind them 💕

All we can do as a community here is continue to push for authenticity and honesty - and encourage people to post their own works, no matter how rough they believe them to be. I'll make more of an effort to reblog fics I love if it helps, and celebrate the thriving creativity that still remains in this space.

I've said it a million times by now, I'd rather sit here and read bullet pointed thoughts/fragments of a story than the same regurgitated nonsense I keep seeing people paste from chatgpt.

Also, the obvious would be running it through an AI detector, although this can be subject to error as with everything else, my above points are kind of to make you think without having to copy and paste fics into the below, but they're always useful too!!

UndetectableAI, Grammarly, ZeroGPT, QuillBot

And I’ll follow on by saying, if you’re using one of the above, use a couple. I ran a text that I knew to be AI through QuillBot, and it said 0% was AI generated, but ZeroGPT said 94% of it was.

[biggestqiblifan]

Why does it feel like Copilot AI would a great tumblr moot???

[quaranmine]

On Wednesday before I gave my presentation I confessed to a new employee that I was worried it would be too long and she brightly told me her life hack was to just let AI rewrite things for her. She said I should put in all my talking points and ask ChatGPT to give me a five minute exactly presentation. I was like....how is the most polite possible way (since this is a new colleague I shouldn't get off on the wrong foot with) that I can express that I will Not be taking this advice. Ever. I told her that I didn't think we were allowed to use ChatGPT at this job (we most certainly are not, it is a nightmare for any type of protected information) and also that I prefer to write all of my own work. Despite my best efforts the last part of that was still passive aggressive, lol.

Something about being a writer makes it so that it's almost offensive to me for someone to suggest I use AI to do my work instead? Like, the day I reach the point where I let AI write something for me is the day y'all need to be checking me for brain damage because clearly I'm losing it

[tsfennec]

... yes, Grammarly, I am Very Aware that I am a wordy person. You don't gotta drag me like this. 😔

[theladycarpathia]

AUTOMATED AI VOICE: Please type your account number. I'm sorry, I do not understand. Please try again. ME: STUPID WORTHLESS MOTHER-FUCKING SHDFKJSDBKJC AI SHOULD DIE. REAL PERSON: Good morning, can I help you? ME: Good morning, you precious bean, I value you for your time and energy ❤️🥰🦄🌹

[mayonaisalspray]

Thinking about a few days ago where I was complaining about really not wanting to do an assignment because I had been working all day and was told that I should just use chatgpt to do it. And I had to be normal about it lest I become the weirdo in the situation

[winny-the-waggishly-werid-wizard]

Anyone within the ages of 18-26 wanna talk to me about books movies nature idrc just want to talk to internet strangers and tumblr has the best internet strangers.

[azumetapraline]

i wished my fic could just

write itself

but alas i am a creator, and i must create.

[bmpmp3]

My sudden onset and inexplicable obsession with genbu in the past year is really fun because I'm developing a lot of tuning skills that apply to literally no other voicebank and will never apply to any others again

[gutsfics]

tbh i think that the "is ai generated content art" argument is roughly the same as "is Duchamp's Fountain art". i think that the argument of if its art is what makes it art, because it is invoking some kind of feeling. starting a conversation. if that makes sense?

that being said, absolutely there is a time and a place for it (ie not replacing the work of paid artists). and it should only be trained on artsts who have consented to it.

i dont want to read fiction made by ai bc if you didn't take the time to write something, why should i take the time to read it? and for non fiction, it'll be riddled with incorrect facts because an algorithm "decided" that this word is the most likely to go next in the sentence based on what it's database says, even if its factually incorrect.

one of my favorite pieces of art is an ai image that was sent to a company that makes paint-by-numbers, and then filled out by a person. the artist, tumblr user rigatonidanza, asks "is this art? at what point does it become art? can it never be art because its ai, or because its paint by numbers? is it art because it elicits a response from the viewer?"

the point of art is to make you think, and to make you feel. does it stop being art if its not made by human hands, the way algorithmically generated images are? is it not art if its something mass produced, made for intentions other than what the artist uses it for, like Fountain? even if looking at it makes you feel something? anything? even if that "something" is a negative emotion?

all that said. if you use generative algorithms as an excuse to not pay real artists, you are a human tar pit. may you always immedeatly stub your toe after it stops hurting from the last time you stubbed it.

[nbstevonnie]

so like. i'm not as hardcore anti-ai as it seems most on this site are. but i am BAFFLED at some of the things people irl will be shocked that chatgpt and it's ilk can't do. they'll say things like 'i asked it to write me a logic puzzle but the answer didn't make any sense :(' or 'i asked it to tell me the relevant bit of law for the situation i'm in but it was totally wrong :(' or 'i asked it to make me a recipe from the things in my cupboard but it tasted disgusting :('

like. no shit!

as soon as you realise that current ai is literally just an algorithm trying to fit new data to the limited number of examples it's been shown, i.e. trying to work out what "most likely" comes next, you know that it just cannot do these things.

of course it's going to give you a logic puzzle with no logical answer! it knows what a logic puzzle sounds like ("a man lies dead on the ground, no footsteps nearby...") and it knows what an answer sounds like ("he died of a heart attack"). that doesn't mean it knows how to connect the two.

of course it's going to tell you a law that's completely irrelevant or just made up! it knows what a law sounds like ("an extension of time of three months is available upon request"). that doesn't mean it knows the law or can apply it to a specific set of circumstances.

of course it's going to give you a recipe that doesn't take taste into account! it knows what a recipe sounds like ("1 tablespoon of rosemary"). that doesn't mean it understands how an ingredient or the amount of it affects a recipe.

the thing that's frustrating about this is that if you're aware of this, you can use ai for things it will do fine at. ask it to write you a speech. ask it to suggest something to eat. ask it to write you the pros and cons of some piece of technology. it will give you enough ideas to tear apart and re-use as your own just fine with any of those prompts. ask it to actually be creative or provide real-world advice? a fool's prompt.

[stakdai]

My online bio was boring. So I let an AI roast it. 🔥

send help. i am not okay. 😭

i let an AI roast my online bio and i am PERSONALLY ATTACKED. 💀 it told me my profile pic looked like i was ‘trying to solve a simple math problem’ and that my bio was giving ‘i have no personality but i do own a plant.’

I'M SCREAMING.

But like... after I wiped away my tears, I realized... it was right??? My bio was so boring. But now?? Now it's ✨amazing✨.

Seriously, if you're trying to make your dating app profile less ‘generic straight person’ and more ‘actually interesting queer person with hobbies and a personality’ you HAVE to try this. It’s the brutal honesty we all need.

I wrote a whole thing about my descent into madness for the sake of self-improvement. It's a wild ride. Check it out if you're ready to get absolutely humbled by a robot. 🔥

Read the full, hilarious story here: https://stakdai.com/roast-ai-the-ultimate-guide/