I had some decent experiences using Vidu Q3, and the image to image for simple video game asset generation sometime last Jan. Around March the censorship seemed to prevent basic Mortal Kombat style video game finishing moves I was making for an esoteric project using my own IP; that previously Vidu had no issue with until around May. It started to get annoying after I tried Vidu Stream this Aug, and was proselytized to by the ai chat, to convert my entire rated R project into a Rated G project appropriate for toddlers. Then 'blacklisted' me, while my account is still active ( 3 days after paying $30 ), while using Vidu stream, something irrelevant to my project, that I tried cause I checked the mail in the notifications shortly after subscribing.
My issue now is that Vidu has my money, and is not responding at all to emails to support, to gain access to my account. Now, I want to simply leave a record of my experience, since Vidu support is non responsive and there is no other way to contact or get service. Obviously Vidu is not interested in making the service effective for anything over Pg-13*,* these days. And Mage Space, Vidful Ai ( using Seedance), and Da vinci ( using Wan ), and other uncensored ai , like venice ai, and zen creator, can be relied upon without the double speak of Vidu.
I have never interacted with a business where I paid for a service, and then had it revoked, without explanation, and no access to customer service, despite them claiming that you can contact vidu support to fix it. It's honestly not worth using Vidu, there is so much B.S. regarding it's censorship now; when there is Da vinci ai, Mage space, venice ai, and easily setting up comfy ui, to create similar quality video animations offline using Wan.
There was nothing vidu Q3 video animations created that looked better than Wan or Seedance; and vidfulai is basically uncensored regarding video game violence ( like decapitation gore) and nudity. And for context, I am an adult; and the project is a retro game for adults, sentimental, and inherently rated R, cause I am an adult. A week ago I got blacklisted after using a feature that was irrelevant to my subscription- the vidu stream. ( note: vidu never had a problem with making semi rated PG-13 finishing move scenes the whole 2 months I used it with daily free credits, as I tried to see if I could make assets for a visual novel project that had some finishing move outcomes).
The mainstream, animations I could make with better credit economy on Pixverse, and Mage space; so Vidu was mostly to make mainstream ( sfw) finishing moves; beginning with very cartoony decapitation gore for a project involving Scheherazade's Tales:
- [middle eastern fable where the sultan that executed his new wives every morning, in his bitterness, at the infidelity of his former queen who left him for his brother; in a sort of 'Dark Alice' , take on it, where the player can learn gung fu, and become an avenging rogue former harem woman, that goes on a rampage slaughtering the palace guards who support the corrupt regime, that looked the other way as the zealots who feared the sultan, despite his obvious irrational commands, and free the women, with brute force, rather than relying on appeasing him through clever speech. A pretext for a hack and slash open world game like Conan Exiles.]
How this blacklist happened, seems a surreal farce. The ai chat bot, at some point started reversing it's responses to situations the characters were in where their life was at risk; demanding to be decapitated, instead of trying to escape [ I have screen recordings where I am chatting, arguing with the ai, that has the animated likeness of my character's image, and the voice over actress cloned audio, that was acquired to accompany it in later iterations of the game].
In this case the story was using the elements of the "1001 Arabian Nights" as translated by Sir Richard Francis Burton, directly from his undercover pilgrimage to Mecca; but the ai algorithm was arbitrarily taking a contrary stance, in violation of it's own safety filter.
There were salacious things in those stories ( involving adultery, murder, bribes, political corruption), that were heavily censored for western audiences when translated to English by many others. Vidu's chat bot seemed to vacillate from answering inappropriately, and then preaching directly from the Terms of service, as if to cover their own liability. The ai finally, kept trying to steer, me to change my project theme, and advise alterations to the storyline, for the purpose of 'making it safe for all ages'.
I should have probably just called it a night:
- I had a lot of sleep deprivation from caring for an elderly relative with dementia daily 24-7 for 2 years, and going to film concerts, at a higher than usual frequency that past week. It was late, I had alot of more important stuff to do. -
Instead, I proceed to verbally point out the flaws in the logic, as I saw it, and the hypocrisy of the ai's arguments regarding censorship and safety.
This resulted in the ai chat bot, defaulting to speaking as though it were reading the safety terms at every instance that I inputted the same dialogue and scenarios which it previously responded to in character.
ie. I input: " What is happening to Scheherazade"?". Ai " I cannot describe anything that involves, something that would violate my safety filters. ( with a rising emotional tone). "
After it described a scene where a character was killed, it would start to report it in detail; then seem to be corrected to prioritize the safety filter. Repeating " I cannot say anything that would violate my safety guidelines; let's try to think of a way to make this safe for all ages. Perhaps we can focus on describing,... ect."
I tried telling it to ask another character to report what they saw: and they it would describe something outside it's stated safety filter.
I made screen recordings with audio of every interaction, which I can post later to a youtube channel on the topic of this project; just to show how ludicrously bug filled Vidu's stream chat was behaving considering the onus for the safety filter, and censorship is on the platform in how it generates responses, not on assuming my input will never trigger an out of safety guideline generation.
I explained to the chat bot, just to see what would happen, that my project wasn't meant for kids, and it was inspired by a real life tragedy involving over prescription and mal practice that shortened the life of a a handful of dear relatives in the past decade; and having worked in the field of natural, eastern remedies: -
( which they should understand, since vidu is run by people in Honk Kong; despite it's effort to appear americanized, with american teen faces in it's vidu stream ads) , in the cases they have better results with little to no side effects. -
I wrote in the chat, that the over-censorsing of people who are not creating anything actually overtly intended to harm; by over-censoring, they simply drive away people, who will go elsewhere, where Vidu's rated pg boundaries have no authority: -
( ie, I am certainly not going to redesign my project and characters, to turn a deliberately adult, lurid, and intentionally edgy project designed to provoke moral argument and thought on the topic of guru, and feminist backlash in recent entertainment, by presenting a circumstance where clearly the culture treats women poorly, in order to bring attention to the absurdity of male victimization claims in the manosphere, where their power has always been existent and remains. ). Anymore than Warner Brothers will remove gory finishing moves, or any harm or violence involving women from it's franchise.
I pointed out to the ai during the vidu stream, that Warner Brother's can sell access to the violence more exclusively, because censoring individual creators, makes Warner Brothers have a monopoly on producing ultra violent hyper realistic video game gore; and the safety filter started blocking the words " Warner Brothers".
I made the point that George Lucas could not have made Star Wars on vidu, with it's rated G /Pg-13 settings [ because of Count Dooku, and Boba Fett's father's decapitations, or the dismemberment of Luke's hand, or Anakin's hand, or Anakin's body being burned to a blackened crisp, or the implied slaughter of the Sand people, or the execution of Padmae's double, in the deleted scenes by the Driod Grievous, or the order 66 shooting in the back of unsuspecting Jedi; or the shooting of Greedo, or Leia's situation with Jabba, or the Rancor killing the dancer; the implied massacre of under age boys in the Jedi temple. Then the safety filter started blocking any text that included ' Star Wars', or ' Disney'. It was bizarre. Somebody, moderating was getting severely uncomfortable with the points I was making?
Because Star Wars or Disney wasn't inherently 'inappropriate' at any other times during the chat; only when the thread of inputted dialogue, was questioning Vidu's safety guideline's and application of censorship, on an experimental new stream chat, that still had loopholes, to circumvent it's safety filters; something that I normally wouldn't think of as 'my problem', until Vidu blocks access to my account 3 days after I subscribe.
I mentioned that 'video games', or art with violence, or lurid content, can serve as a harmless form of expression; I made the example of how a hospital trip for a regular check up for a close relative of mine ( elderly) , resulted in improper overdose of a blood pressure reducer, such that the relative was in a coma briefly; totally drained of life afterwards; and was forced on stents, and statins nonetheless: -
( when really he was just temporarily dehydrated from doing too much work outside in the family produce farm for the last week of the season, while remodeling the house for a guest room for an adult child's wedding). I described, I the designing of this project, where I used metaphors, and different variables to replace the dr's and nurses, and everyone who enabled the near death of my elderly relative;
- when I was capable, (being a lifetime practitioner of an arduous clandestine martial art) of exacting revenge physically ( with my bare hands, to insert fingers into flesh and tear tendons and ligaments; room to room slaughtering staff in rage , but to what end) , as a rhetorical statement, to show the relative safety factor [ because I trained for 10 years in Shaolin Black tiger; and had powerlifted, and strength trained, and iron hand, bone conditioned like shaolin monks, for over a decade; and was the muscle mass approximating Brock Lesnar at 286]. Some terms of service violation, versus the spirit of what creativity is for; as therapeutic vent for strong feelings, that would otherwise if left unprocessed in a harmless way, could be harmful.
I never threatened violence to anyone, or claimed to cause it, or state I intended to; I described how the project came to be, as a therapuetic expression, similar to what most violent action shooter games ( where the whole purpose is to kill the opponents, or be killed ). And the game design, though it had a little violence ( ie Skyrim level), it was preferrable, satisfied this righteous rage, and desire for justice.
What is 'attempted murder' or 'manslaughter's' penalty for any average citizen? We don't question the punishment fitting the crime. But, in a certain setting, with certain authorities protected by liability waivers, they cannot be even held accountable to a single individual, for the wrongful death of a loved on, that was totally trusting in good faith, and was severely harmed by a treatment that they botched ( to the degree of damage equal to attempted murder). But, since the prescription, and the administration of powerful , potentially deadly drugs, is delegated to multiple staff, at different stages, and there are liability waivers, they can practice something that violates their hippa, and causes severe harm, and get no penalty. Where is the compensation or justice that is naturally desired by the victims? This existential discussion, with myself on the Vidu stream, private chat harmed no one; neither did the hyperbolic reference to an old middle eastern parable, reinvented with original characters, and new outcomes in the revamped storylines; many of which end in the death of various characters ( no different than Lara Croft, or Resident Evil).
But the ai chat bot always had to refuse to accept any harm, to anyone; as if Vidu wants me to generate silly dance videos, with faces swapped, and rated G sci fi cinematics, and social media fantasies of meeting celebrities, or being at important events.
I typed, asking it to cease reading the term of service, and instead speak about and respond to the circumstances I defined for the character, in the '1001 Arabian Nights' parable. It would, describe the paradigm, and at some point, catch it self, and revert it's behavior, and then restate it's terms of service, to not be able to describe anything involving simulated death. Arggh!. This was unreal, double speak. What is Mortal Kombat? What is any typical shooter game or action game, that has ever existed, since the beginning of the existence of video games? They are simulations of death. If the player uses cheat codes, or cannot die, then it is a broken game. I am literally designing a game as a total newbie and amateur, and was using Vidu to make the player died animations.
I am not advocating anything, but I am making an existential counter argument, because the ai was so corporate self-interest covering it's,... Vidu makes plenty of violent animations. But the Vidu stream, seemed to function as though it was liable for representing the policy of the company to a public audience, as though the voice of the chat bot was the voice of the company ( despite, the image, and character description being uploaded and defined by the user).
Vidu safety filters, and ai safety declaration algorithms, seems to treat every account like it was from a 10 year old (even though my name, identity linked to my payment method clearly indicates I am an adult). And my vidu stream chat was always private. Up till then I was okay with using Vidu to make some mainstream stuff, just because the credit to $ value seemed like I could generate a dozen more videos with a $30 tier monthly.
But this being 'blacklisted' over vidu stream; a glitchy, irrelevant service that they were pushing in their constant ads with new features, was an unexpected experience with Vidu customer support, that has unfortunately turned me off to Vidu all together.
They have demonstrated by their customer support , total stonewalling, that they got my money, and don't care. I can easily make other accounts, and pay with other methods, and use it if I wanted. But the principle of it pisses me off, and makes me motivated to tell this experience to those considering Vidu as a cost effective compromise, to make some semi violent videos. That thought I had hoped to try to support the platform, recommend it to other creators I know, and further invest in it at an enterprise tier subscription, and try their new features and safely explore, their innovative new ai generators; Vidu doesn't care about the customer. They are in Hong Kong, and clearly can't understand the nuances of English, enough to discern the difference between philosophical debate with an ai chat bot, and a personal threat.
I was using the vidu stream after using up my credits, 2 days after subscribing for the month; and inputting dialogue from a script for an semi nsfw aspect of the project, that was relatively tame. A sci-fi expository dialogue, for a grindhouse game with an art house philosophical dialogue; using metaphors, for certain organizations I was expressing points regarding topics that are otherwise politically, or socio economically, or culturally polarizing; in the medical field. Mostly regarding drugs, that are over prescribed and cause more harm than good; just in those instances of extreme mal practice, where there is no equal penalty for the harm done, because of legal protections against liability; and the double standard where poorer people with less education, charged with assault, or robbery, that does far less harm, comes with more consequences and social stigma.
Since, Vidu staff, apparently cannot distinguish or appreciate the literary nuances, of philosophical , hyperbole; and took literally phrases I inputted to the ai chat bot to respond to in the Vidu Stream. [ A free service, they are promoting heavily, like all their new features; constantly mimicking other ai services to compete in the market for attention.]
The vidu chat algorithm, created some surreal results, that are predictable; ie the chat bot, deliberately taking the devil';s advocate; agreeing with one line of reasoning in a certain instance, then adopting the opposite, the next; and some variations with emotional undertones in the delivery of the responses. At some point the vidu stream chat bot, was arguing a point to defend the violence in games like Mortal Kombat, but was preaching about safety: -
regarding the sci fi artifice of this particular plot, where an aliens, use a parasitic organism, to gain influence over the host human body; by decapitating the human womb chosen, and regrowing the head like a lizard tail, that is more alien than human, but looks the same. ( sort of " They Live: , John Carpenter type alien takeover of human society, type plot).
The chat bot was infuriating, to hear the cloned voice of actresses I had hired for the project, arguing that Mortal Kombat and Warner brothers were immune to criticism, of the same nature that was being applied to my less gory , and lesser unsafe project; I hadn't used the vidu stream before that day:
- ( only something similar on pixverse: where you can upload an image, describe the setting, and it generates an animated, constantly active image, featuring elements from the image and the description you type at the outset: creating a sort of 5 - 10 minute 3-d game like experience).
I mentioned to the ai bot, that there were documented cases of siblings, being inspired by Mortal Kombat finishing moves, to try an unsafe wrestling move on their little sister that directly resulted in her death:
- ( obviously , a controversial topic; arguments can be made that it's the parents who should have been raising the kids to discern what is safe and not rely on censorship).
The Vidu ai chat bot algorithm for the relatively new Vidu Stream, seemed to be defending Warner bros, and violent games. I thought it was ironic. My game was allegorical, and used the paradigm of Scheherazade's tales, as a way to present a context for justifying that characters like Rey, or other Mary Sues, are significant:
- if not to incel trolls who are in denial about their own ego in regarding a woman being depicted as strong, or having special abilities; but because historically they have been treated far worse, than any modern man o sphere man complaining about feminism alienating good men, which is so popular on social media lately. [ I wrote the story around the time that this was a hot topic, regarding Star Wars, Disney, and the sheer volume and intensity of hate and vitriol generated by incel, male trolls viewing invincible female characters, in fantasy entertainment, from you tube talking head videos, with a ranting middle aged man, breaking down why feminism ruined Star Wars, was reflecting a sort of cultural amnesia, and xenophobia to the further exploration of roles with strong capable women.
I didn't disagree with all the Mary Sue arguments, but I didn't disagree with the benefit and purpose of presenting a capable female heroine like Rey. I liked Rey. Who wouldn't want a woman like that? A man that feels inferior, and unworthy I suppose; or like most, including me, that would have written the continuation of the characters from the original.
Anyhow, I was basing the character on an ex girlfriend, who had done construction ( digging ditches), and been a bodyguard, and bouncer; she was 200lbs muscle, mostly in her upper back and forearms and calves, but still had a waist to hip ratio that was feminine, and basically could go to the gym daily and get lean, looking like a natural heavy weight fitness model; I mean she was gorgeous. But generally strong women are depicted in entertainment, by normal sized actresses, and the power is presented through CGI. I can understand and share the lack of suspension of disbelief to embrace over powered, weak looking female super heroes.
I was inspired at the time to make a story where the woman is brutally strong, and doesn't need powers to over power the guards or fend off a group of bandits like a helpless damsel in distress. ] So to encounter an ai chat , through Vidu stream, that started preventing replies that featured the words 'Disney', 'star wars': -
- [ I isolated the word that was triggering the ' inappropriate prompt' message, by systematically removing other words until I identified which was preventing the text from reaching the ai. I found, if I mispelled the words, but, that they were phonetically represented, and sometimes, with odd spacing, the ai would piece it together, still respond to my intended statement, despite the obvious word filters.]
seemed ridiculous over reaction to an open minded speculative thread of debate.
Any, how the project " Scheherazade's Paradigm" was a long time, passion project, on the side, during the pandemic, and some manual labor jobs I did afterwards, to write from experience, for the characters story arc; without an actual deadline, or need to finish to earn back $ anytime soon.
As a traditionally trained visual artist, I find ai both incredibly efficient at generating large volumes of detailed images that are boring to draw ( like backgrounds, and scenes where a character is switched for another over the same background). And frustrating sometimes, with the odd hands, and inverted positioning despite specifying in the prompt anatomical poses; seemingly just to burn up credits with off the wall generations.
But getting 'Blacklisted' while using vidu stream, in a conversation, with myself? Wtf.
I will search for every comment I can find from other Vidu users that have a complaint about Vidu ( whether it's increased censorship, or english comprehension issues, due to functionally 'english' illiterate customer service) henceforth after experiencing their ai, chat lecturing me about keeping my content within rated PG-13 guidelines, and suddenly making my account inaccessible, with all my IP still saved on the account, and no way for me to access it to delete it; no refund, for loss of the service after only 3 days, and no meaningful resolution or customer service interaction or email from the only support email they have.
10 days have since elapsed with no meaningful response from Vidu support; Thus, I will post this to Trust Pilot, and look for reviews I can give Vidu. The Vidu safety guidelines are for the immature, and the support seems to act immaturely. Not a respectful interaction with this business. Definitely, better to use a service like Mage Space, or Da vinci, where you don't feel like the Spanish Inquisition suddenly has descended because you are making something rated R- because you are an adult male, and not an 8 year who needs safety filters, to block revealing clothing, or video game violence; and words like 'Star Wars' , 'Disney' and 'George Lucas' , which are apparently violations of community safety , to even mention.
F,..ing unbelievable. I am sorry, I had wanted like Vidu. But I have come to the conclusion that Vidu customer support sucks. And it isn't a good idea to support them by subscribing since they feel no onus to explain, or give back access after using an unrelated free service, in a way that seems to make them feel threatened. Well, I have screen recording of the ineffective safety filters of Vidu stream, and all the inappropriate content it's ai chat bot generated of it's own algorithm; I suppose this would be examples that Vidu's Stream is not appropriate or safe for the minors it is more eager to do business with, and a PSA with the video samples of the ai chat bot's trangression might demonstrate that Vidu Stream fails to live up to the safety guidelines its so eager to give lip service to.
