Fast forward a few months, I’m happy to see Ukraine has now been whitelisted as a keyword after testing it out again today. I am however concerned by an email I received from a Ukrainian person who read my original thread. Here’s what he had to say:
“Hi, I’m writing to you because I saw your thread on OpenAI forum. It’s actually even worse: Ukrainian users can’t even register (which is why I’m writing you here because I can’t register). It’s among banned countries, and here’s the list of them: China, Russia, Ukraine, Afghanistan, Belarus, Venezuela and Iran. They put Ukraine in this list of terrorist countries. I hope you can rise this problem on forums since I can’t.” [end of quote]
While I can understand why most of those countries are on the banned list, I see no reason for Ukrainian residents to be discriminated against. They are the victims of Russian aggression. And while I try to stay clear from political / controversial topics, I think it’s pretty safe to say most of the free world supports Ukraine in these trying times.
If all of a sudden major corporations like Apple, Amazon, Microsoft, etc. started sanctioning Ukraine as apparently Dall-E has, there would be a huge uproar. I might even speculate that some of Dall-E’s sponsors would be appalled to learn that Ukrainian residents are banned from this platform. Especially Microsoft if I can use one of their recent blog posts as a reference: Extending our vital technology support for Ukraine - Microsoft On the Issues
That said, I’m not sure how many people will read this. Guessing most people use Discord for sharing images and discussing Dall-E. I can only hope this thread will catch at least one eye from the Dall-E developers so that Ukrainian residents may be allowed to register.
If not, I might look into news or social media outlets for shedding more light on this unfair ban. Please note I don’t want to do this. If you look at Trustpilot, I am the first (and so far only) person to give an overwhelmingly positive review for Dall-E (back in September). I have nothing to gain by casting shade at this point, I genuinely consider Dall-E as a game changer in the visual arts, and want to see it succeed. But not at the cost of marginalizing a whole culture of Ukrainian people who did nothing to deserve a blanket ban.
This goes in line with Musk using his power over at twitter to remove the Ukrainian country code so Ukrainians with 2FA can’t log into their accounts.
It is clear that he has a position of bias against Ukrainians, and he is using his wealth and authority with companies and orgs to push that agenda.
Why OpenAI ban Ukraine if they just can ban keywords for ukraine war topic? That seems not logical, Ukrainians still exists, and still creating content, art… Why Syria not in ban if we will use this logic? Why we should ban things that overload the social networks, for example, people of Ukraine just wanna live, and still exists on this planet, AI will still be used for bad purpose in any chance.
Last move from the Twitter for removing +380 code is awful.
OpenAI is a private company and they can do as they’d like.
If you spent billions of dollars developing an AI and you were a data scientist that didn’t want your AI’s bias to shift towards political topics - you would do the same thing.
Your point about Microsoft, Facebook, etc banning anything to do with Ukraine is invalid because those companies are not training a public facing, consumer-end, generative AI tool.
I’m not sure if this is true. If OpenAI was a typical company, it would be, but they’ve intentionally shouldered a lot of responsibility for Safe and ethical AI as well as the idea that AI can also act as a public good for society at large. It’s built into their very corporate structure since there’s a cap on profit taking for shareholders and the ability to funnel profit into other activities that contribute to society. So given that’s the case, they have made the decision to steer away from particular topics, and those actually may be sound decisions given their goals and mandate. However, because of the larger responsibilities they’ve intentionally decided to shoulder I think it also means that these decisions are open to public discussion. In fact this is probably the single best time for nations, institutions, and others to engage in a discussion and dialogue about AI and where it’s going, and how to nudge it along safely so that it may benefit everyone.
Can we generate a document explaining the conflict in detail and send electronic as well as paper media to the people fighting to see if they wish to make a informed choice to keep fighting or to withdraw their consent. Could you use a team and ChatGPT to do this?
No disputing OpenAI can do what they like, but they did set up this forum meant for discussion. I would hope they welcome constructive feedback for the purpose of bettering the platform, and my criticism relates to ethics as to how they treat Ukraine.
First I noticed a ban on the word Ukraine itself a few months ago. Now Ukraine is allowed in prompts. It’s not a stretch to imagine they can take it one step further by allowing Ukrainian residents to register on OpenAI.
We’re talking about real people here. Many of whom suffered for simply wanting to live free from oppression. Might be easy to lose sight of in the context of AI or in this age of social media where people are often treated like statistics. So while it may be “inconvenient” for Dall-E that the word Ukraine leads to political topics, let’s not forget the more human and compassionate side of the equation. I believe this can be improved upon.
And while OpenAI isn’t a publicly listed company, I think it’s in their best interest to develop a product that sells. Part of that includes listening and reacting to its customers’ suggestions and criticisms. From what I can see, Dall-E isn’t the only platform of its kind. It doesn’t have an impenetrable moat to shield it from competition.
More broadly speaking, I see this as a weak point for OpenAI. I’m looking at Trustpilot again, only to see my review is still only the positive one yet. Others are complaining about lack of support and privacy issues. Might not look like a big deal now, one might say OpenAI gets a pass for how new and disruptive the technology is.
In time however, reviews like those on Trustpilot will matter a whole lot more as consumers weigh in which AI platform to choose, and ultimately what to spend their money on. Consumers who are at the very core human, and as a human consumer who incidentally spent well over a thousand dollars on Dall-E credits, I think I’ve earned the right to give my two cents on the matter. And one of the most glaring issues I see right now is the ban on Ukraine residents.
This is a dumb move. It’s still pretty easy to use it with a simple Poland VPN and a drop-out number. All they get by restricting is generating more shady actors behind VPN and one-time numbers. It’s obvious that you don’t have nationality in the net, it was supposed to be over it. How would you prevent bias shift towards political topics by banning certain countries? This makes zero sense, since ALL people in the world can and WILL be shifting towards political topics. The real reason seems far from everyone’s understanding here
Again, just because you believe that ALL the people in the world WILL be shifting towards political topics does not mean that Open AI has to allow this or that other people share your beliefs quite frankly.
If you do not like the fact that Open AI is trying to protect their neural network during the very fragile stages of development and public testing, then go use a different AI or wait until they unban sensitive topics.
I find people who oppose the political and sensitive topics ban, typically do not care or do not care to understand just how sensitive a generative AI network is. There are 1000’s of legitimate organizations and companies that use OpenAI to generate useful text and images that have nothing to do with political topics. In the current state of the world, if you allowed political topics in the OpenAI system you would pollute all of your other data because of the overwhelming amount of political input that would be going in.
Again, people that oppose this ban will simply say, “Yeah but they shouldn’t do that.” And I’ll tell you that you lack a fundamental core understanding of what it takes to build something like OpenAI. You will definitely be able to ask it political questions one day, just not in the infancy when the data sets are still being trained.
I am a Ukrainian and I am shocked that OpenAI banned Ukraine as if it is a dictatorship or terrorist state similar to Iran, Afganistan, N. Korea and Russia. No words left. This is an unfair punishment for nothing.
You substitute the concepts. From your POV, OpenAI banned the whole Ukraine people from the service because they want to preserve AI from politics. But this is just a nonsense. 8 millions of Ukrainian refugees are in EU now. So, why not ban all EU countries to stay safe?
Or do you think that the main goal of Ukranians now is to feed AI with political sh*t rather than fight against russian invasion and just stay alive?
Also, it would be good to know how many requests from Ukraine were rejected in the last 6 months. I bet this number is less than 1% of the whole Ukraine population.
To me, banning Ukraine because of the possible political discussions is unethical. If this is true, then OpenAI itself fails to be ethical and its well-intended goal may be biased.
But I think that OpenAI’s services could be used in malicious ways on a war front and this is definitely not desired.
Personally I’ve been using Ukraine related news articles to test various functionalities.
For example to distinguish types of content between opinion and professional journalistic content.
Or to see how GPT-3 acts when is presented a true event that includes violence description.
However, there are many Ukrainians who had to flee in Europe due to the war and now they can access it from there, so saying that there’s something against Ukrainians it isn’t right.
OpenAI could make a public statement on this matter, but from the PR point of view may not be desired to attract negative attention.
Alright, think about it this way: ip address blocking just makes little sense, since anyone in the world can be using a private network located in another part of the world, this is cheap, and this is available to anyone at the moment really. So the only thing which clearly draws attention is the list itself, the list of countries and their collocation in the same list, and truly it doesn’t help to actually prevent access to openai, but creates an embarrassment stinking of national discrimination. Everyone in Ukraine will be hating this unfair move, which doesn’t really help company with anything. lose-lose
Of course, I will be intentionally feeding OpenAI all kinds controversial, sensitive and political topics to it, pretending I am funny Ugandian farmer in my 40s, just for fun, because this is internet, and i can rent a server in Uganda and do it, we must test if it will withstand the real world. It is so funny to hear something PUBLIC to the WHOLE INTERNET but a little angry list of coutries will be protected from anything at all, it is just not how it works
PS.
Sending this comment networks packets from Australia, at home in Kiev, which doesn’t make me an Australian
Here is a quote from my old comment - “I find people who oppose the political and sensitive topics ban, typically do not care or do not care to understand just how sensitive a generative AI network is.”
I thought that this was purely about the “topic” of Ukraine, I had no idea Ukrainian people were banned. I am so sorry about that, please accept my apologies!
I will not edit my previous comment, I’ll leave it up there so that way everyone knows why this conversation is the way it is.
But when I made those initial comments, I had NO idea the actual people from Ukraine were banned from using the app.
I thought that Ukraine as a topic was simply banned because it is political and they don’t want that to interfere with their systems.
I had no idea the actual humans that live in the area of Ukraine, were in fact banned from OpenAI. Again, I thought it was just the topic of Ukraine.
I agree with you and I also think that this is wrong and is a very weird rule, especially with no explanation. It is very “dictator” like.
Got it, yeah it would be weird if we all would advocate for ability to use the topic of Ukraine not the silly ip ban itself, which we are all discussing here after all, but following on your topic, which is way more interesting to discuss: In my opinion if there is something that sensitive, then we must be certainly testing it’s borders, and if it’s sanity can be compromised, it should not be released to public
The ip ban is easy to avoid, but with language, it is even easier to shift meanings and overcome any kind of filters to compromise any kind of data system relying on language
Say in this video Luke is revealing different methods of “gaslighting” and tricking the model into giving the answers he wants, so why bother fighting it if anyone will be doing it still? (minute 5) Luke LOVES this AI chat bot - YouTube
As for me, the reason is obvious:
Avoid the fights that will arise after ChatGPT express his opinion about the causes of the conflict.
And avoiding all the slop that will pour out on the OpenAI, who will be made responsible for this political position.
According to the Ministry of Digital Transformation - they were told that this is due to the sanctions have been imposed on the occupied territory of Donbass and Crimea. Unfortunately, they said, they do not have the physical ability to separate the territory of Ukraine from the occupied territory of Ukraine. This is hard to believe, given that others can, but this is the company’s official response.
