This kind of seems like a non-article to me. LLMs are trained on the corpus of written text that exists out in the world, which are overwhelmingly standard English. American dialects effectively only exist while spoken, be it a regional or city dialect, the black or chicano dialect, etc. So how would LLMs learn them? Seems like not a bias by AI models themselves, rather a reflection of the source material.
this post was submitted on 29 Aug 2024
95 points (100.0% liked)
Technology
37573 readers
169 users here now
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
Subcommunities on Beehaw:
This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.
founded 2 years ago
MODERATORS
It's not an article about LLMs not using dialects. In fact, they have learned said dialects and will use them if asked.
What they did was, ask the LLM to suggest adjectives associated with sentences - and it would associate more aggressive or negative adjectives with African dialect.
Seems like not a bias by AI models themselves, rather a reflection of the source material.
All (racial) bias in AI models is actually a reflection of the training data, not of the modelling.
I would assume the small amount of training data written that way doesn't contain that many professional research papers, corporate emails or calm poetry, but would consist mostly of social media posts and comments which have a rather heavy bias towards aggressive and negative.
Seems like not a bias by Al models themselves, rather a reflection of the source material.
That's what is usually meant by AI bias: a bias in the material used to train the model that reflects in its behavior
But why is it even mentioned then? It's FUCKING OBVIOUS. It's like saying "AIs are biased towards english and neglect latin" or smth ffs
I feel like not everyone is conscious of these biases and we need to raise the awareness and try preventing for example HR people from buying AI-based screening software that has a strong bias that is not disclosed by their vendors (because why would you advertise that?)
I was confused how a resume or application would be largely affected, but the article points out that software is often used to look over social media now as part of hiring (which is awful).
The bias when it determined guilt or considered consequences for a crime is concerning as more law enforcement agencies integrate black box algorithms into investigative work.
Great comparison, a dialect used by millions of people to a dead language. It really shows how much you care about the people who speak that dialect...
AIs are trained on what is written in the Internet. Latin is not spoken, it's written. But even then, it's rarely used. African american is a dialect, it's only present in speech.
You need to get out more. I totally get that you would think that’s the case, but only if you’re not exploring parts of the internet outside your bubble. It’s absolutely written.
There are actually quite a few books written in AAVE...the earliest I'm aware of is their eyes were watching god, from the 1930s. The Color Purple, Beloved, The Sellout, the books of Chester Himes...
It’s FUCKING OBVIOUS
What is obvious to you is not always obvious to others. There are already countless examples of AI being used to do things like sort through applicants for jobs, who gets audited for child protective services, and who can get a visa for a country.
But it's also more insidious than that, because the far reaching implications of this bias often cannot be predicted. For example, excluding all gender data from training ended up making sexism worse in this real world example of financial lending assisted by AI and the same was true for apple's credit card and we even have full-blown articles showing how the removal of data can actually reinforce bias indicating that it's not just what material is used to train the model but what data is not used or explicitly removed.
This is so much more complicated than "this is obvious" and there's a lot of signs pointing towards the need for regulation around AI and ML models being used in places it really matters, such as decision making, until we understand it a lot better.
Yeah this seems like a non-issue to me as well; the source material for the models is probably the cause of this bias.
I also don't think there's a lot of sources for this manner of speaking. Let's also not forget that there's oftentimes instructions given to the LLM that ask it to avoid certain topics which it will in fact do.
I'm from the Midwest US and I know there are words and sounds I pronounce with a Midwestern accent but I can still type and spell them correctly.
If'n I typ lik dis den o'course people gonna think I hev the big dumb or that I'm a mole from a Redwall book.
All the people here saying "well of course because they weren't trained on AAVE":
THAT'S THE WHOLE POINT
It's the same reason facial recognition and voice recognition software have a difficult time with anyone who isn't white or a speaker of perfect, uninflected standard english. The bias is created by the developers, conscious or not, because they only train it on what's in their own bubble. If you don't have diverse teams behind the development and training, you will create this bias, whether you want to or not. This is well known.
There's also just the issue of the fact that there's significantly more books, articles, etc. written in standard english vs AAVE so that's gonna be a huuuge barrier to overcome regardless of diversity of development and training teams. Not to say diversity isn't important, but also that there's just certain challenges surrounding finding adequate amounts of high quality training data, especially for less mainstream concepts. It's the same reason an AI couldn't give a summary of a book that has almost no info abt it on the internet.
The problem is that they trained the models using millions of pirated books in standard english.
AAE is mostly used when spoken: they also pirated also millions of tv series and youtube videos that can contain that, but as of now, it was mostly for training voice recognition models
(proof that they pirated television content and youtube videos to train whisper: https://community.openai.com/t/subtitles-created-by-amara-org-qtss-etc/462561 - https://gist.github.com/riotbib/3b3c5f817b55b68801d14b8bdb02df09)
Given the responses in this thread, it seems that the same bias exists even in ostensibly leftist spaces. Yikes.
Y'all need to get out more.
It'd be interesting to see how much this changes if you were to restrict the training dataset to books written in the last twenty years, I suspect the model would be a lot less negative. Older books tend to include stuff which does not fit with modern ideals and it'd be a real struggle to avoid this if such texts are used for training.
For example I was recently reading a couple of the sequels to The Thirty-Nine Steps (written during WW1) and they include multiple instances that really date them to an earlier era with the main character casually throwing out jarringly racist stuff about black South Africans, Germans, the Irish, and basically anyone else who wasn't properly English. Train an AI on that and you're introducing the chance for problematic output - and chances are most LLMs have been trained on this series since they're now public domain and easily available.
I don't like the idea of restricting the model's corpus further. Rather, I think it would be good if it used a bigger corpus, but added the date of origin for each element as further context.
Separately, I think it could be good to train another LLM to recognize biases in various content, and then use that to add further context for the main LLM when it ingests that content. I'm not sure how to avoid bias in that second LLM, though. Maybe complete lack of bias is an unattainable ideal that you can only approach without ever reaching it.
I just tested out the classic "She working" vs "She be working," and the machine got it backwards. It can't translate to AAVE, but it probably can appear to be well enough for people who wouldn't know the difference. In terms of available written materials just by population and historical access it seems like there would be way more incorrect white imitations of AAVE to draw from than its correct usage. Like a lot of LLM issues, it's been a problem for a loooong time but is now being put into overdrive by being automated.