Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
With over 46,000,000 + active user accounts, they deal in over 1,800+ support tickets every day. They are not here to handhold anyone, so they use pre-made canned responses and are limited in what they can do. And the support team does not harass or discriminate anyone.
Just because you have asked them in the past about what you claimed was a broken game, went well above the refund limits, doesn't mean that denying the refund was in anyway discriminatory.
And when they repeatedly closed your tickets about an issue because they could no longer assist you with, that doesn't equal harassment.
With the guidance of a technician of course.
Can you provide more information for the basis of this statement. It is scary that people might actually believe this. The fact AI runs on computer that fundamentally uses logic gates does not inherently make the results of the AI "logical." LLM mimics the behavior of the training data. There are a number of studies that means the behavior replicates the bias that still produce results that discriminate.
https://www.nature.com/articles/s41586-024-07856-5
https://news.uchicago.edu/story/ai-biased-against-speakers-african-american-english-study-finds
https://hai.stanford.edu/news/covert-racism-ai-how-language-models-are-reinforcing-outdated-stereotypes
Oddly, some video games predicted this type of issue. For example, Portal 2 from 2011 has a key AI character that has very negative behavior because of the training of the AI was sourced.
I am not against constructive criticism of Steam support and seeking to make it better. But current hallucination generators do not qualify, in my opinion, as an improvement. Given the same problems description was to be provided to Steam support, OpenAI and Gemini, I still have the most confident that Steam support has the greatest chance to make progress on the issues between the three.
Exactly, Steam AI
Far too many mediocrity forum keyboard warriors who sits around all day, looking for attention and pretending they are relevant.
they could be using ai already, for all we know
we assume that steam went the normal, outsourced route
we could be dead wrong and have been ai'd this whole time
second
i am curious what you think they would be using to train it?
more than likely it would just be the data gathered from past and current moderation
meaning it would not change at all
And neither of those outcomes are bound to improve with AI based support.
I think I encountered it with my contact with King.com (Candy Crush Saga & other Saga games makers) email tech support. I had that a few months ago.
No real help was provided! The first reply clearly said what they needed from me in terms of info & screenshots. Once I sent that all, and I sent it correctly - it kept on repeating previous - obviously useless - reply text over & over & over & over again.
Thankfully, their online forums still have real people reading & writing - including real/live mods, as well as even some of their developers & testers.
So, my issue was then fixed mostly due to the online forums, and not due to tech support. :)
(Unlike the legacy, dinosaur-style Steam forums King.com has modern forums that allow for easy adding of screenshots, in several ways.
You can attach them as files, or put them inline with Copy/Paste.
Other types of files can be attached too, unlike here.
Steam forums can only learn from that, but I guess are not likely to..)