I think a focus on the source of the misinformation is misplaced
It’s the power of that source to generate misinfo at a faster speed and for close to no cost that’s a more pressing issue here.
I don’t think this is particularly likely to happen, but imagine I use a LLM to create legal documents to spin up non-profit companies for very little cost, I hire a single lawyer to just file these documents without looking at them and only review if they get rejected. I could create an entire network of fake reporting companies fairly easily. I can then have a LLM write up a bunch of fake news, post it to websites for these fake reporting companies, and embed an additional layer of reporting on top of the reporting to make it seem legit. Perhaps some of the reports are actually twitter bots, Instagram bots, etc. spinning up images with false info on it, and paying for bot farms to surface these posts enough for them to catch on and spread naturally on outrage or political content alone. This kind of reporting might seem above-board enough to actually make it to some reporting websites which in turn could cause it to show up in major media. This could end up with real people creating Wikipedia pages or updating existing information on the internet and sourcing these entirely manufactured stories. While there are some outlets out there who do their research and there are places which fact check or might question these sources, imagine I’m able to absolutely flood the internet with this. At what point of all total reporting/sharing/news/tweeting/youtubing/tiktoking/etc does this become something which our system can actually support investigating?
I also think it’s important to consider the human element - imagine I am an actor interested in spreading misinformation and I have access to a LLM. I can outsource the bulk of my writing to this LLM - I can simply tell it to write a piece about something I wish to spread, and then review it as a human and make minor tweaks to the phrasing, combine multiple responses, or otherwise use it as a fast synthesis engine. I now have more time to spread this misinformation online, meaning that I can reach more venues and create misinformation much quicker than I could previously. This is also a potential vector through which misinformation can be spread more quickly through the use of LLMs. In fact, I’m positive this vector is already being used by many.
However none of that touches on what I think is the most pressing issue of all, the use of AI outside it’s scope and a fundamental misunderstanding of inherent bias in systemic structure. I’ve seen cases where AI was used to determine when people should or shouldn’t receive governmental assistance. I’ve seen AI used to flag when people should be audited. I’ve seen AI used by police to determine who’s likely to commit a crime. Language models aren’t regularly used at policy scale, but language models also have deeply problematic biases. I think we need to be rethinking when AI is appropriate and the limitations of it and to consider the ethical implications during the very design of the model itself or we’re going to have far reaching consequences which will simply amplify existing systemic biases by reinforcing them in their application. Imagine that we trained a model on IRS audits and used it to determine whether someone deserved an audit. We’d end up with an even more racist system than we currently have. We need to stop the over-application of AI because we often have a fundamental misunderstanding of scope, reach, and the very systems we are training them on.
Unfortunately, AI’s typical problem with biases, in particular those towards certain minorities which are discriminated against online, did not warrant making this release. It only gets a tiny mention under limitations:
GPT-4 still has many known limitations that we are working to address, such as social biases, hallucinations, and adversarial prompts.
This will never end up happening, because big business has its hands in every government, but tracking of any sort really needs to be opt-in, rather than opt-out. In California, for example, this is how it works for companies which like to send out those “we want to share information with our business partners” emails, documents, etc. If you are a resident in California and do not reply, by law the company must assume that you opted-out.
I am willing to contribute storage (I have several TB), but I am somewhat bandwidth limited, so I need to be a bit careful with hosting too many images to not impact the other services that I run on the same connection.
How would you accomplish this? I have plenty of bandwidth and plenty of storage I can subsection as a possible solution (hell even buying a raspberry pi and an old hard drive wouldn’t be all that expensive and potentially a fun project) but I really don’t even have an idea of how to connect this to the lemmy instance
I find it rather interesting that some of the places most keen on adopting AI are some of the places most plagued by racism. Experts in the field pretty unanimously agree that nearly all AI is racist, so choosing a target system that’s already really racist is just not a good idea.
Unfortunately at the end of the day, capitalism is likely to win. This will likely be sold to police departments in the coming months and years, despite this article or any attention it’s going to receive.
This is exactly the kind of AI application that is almost assured to happen in financially strained systems, especially systems of government that are chronically underfunded, that are most at risk of causing serious harm because nearly all algorithms are biased and in particular, racist.
This is the use of AI that scares me the most, and I’m glad it’s facing scrutiny. I just hope we put in extremely strong protections ASAP. Sadly, most people in politics do not see how dangerous using AI for these applications can be, so we most likely will see a lot more of this before we see any regulation.
If you’re curious as to why these kinds of applications are nearly all biased, the following quote from the article helps to explain
The Allegheny Family Screening Tool was specifically designed to predict the risk that a child will be placed in foster care in the two years after the family is investigated.
They are comparing variables to an outcome - the outcome is one which is influenced by existing social structures and biases. This is like correlating the risk of ending up in jail with factors which might loosely correlate with race. What will end up happening is that you’ll find the strongest indicators of race, in particular if you are black, and these will also be the strongest indicators of ending up in jail because our system has these biases and jails black individuals at a much higher rate than individuals of other races.
The same is happening here. The chances of a child being placed in foster care depend heavily on the parents race. We are not assessing how well the child is being treated or whether they might need support, we are assessing the risk that the child will be moved to foster care (which can alternatively be read as assessing the likelihood that the child is of a non-white race). This distinction is critical to understand when AI is reinforcing existing biases.
A few issues I’ve seen with adoption in the federated/open source world-
There is a technical barrier to entry. The fact that you’re on a website that’s connected to other different websites in the same interface is one that people aren’t particularly familiar with. For a social website, questions around moderation and who you’re interacting with are problems which are hard to address if you’re unwilling or incapable of learning the terminology you need to learn to understand how this works.
Each entry point into this website system is slightly different as well - how it presents itself, the design, who participates on that entry point, what kind of discussions exist. You might stumble across a lemmy instance as your first introduction to lemmy that doesn’t appeal to you and not recognize that it’s not everything that’s available on lemmy and discovering that can be difficult. The same is true of other federated websites.
As you mentioned there’s also issues with algorithmic feed. This is what leads a lot of people to stick with a particular platform. They want content to come to them, rather than searching for it, and they aren’t always aware what content they want. Federated content is much more pull oriented than push oriented and until someone codes an algorithm to push I think there will be a lot of resistance with a particular subset of individuals who are interested in pushed content rather than pulled
Their hobbies likely aren’t causing them to have negative feelings, whereas their work more likely is. Humans are somewhat biased towards needing to vent and talk about issues which cause them negative feelings that they have to do.
People also talk about work for a variety of social reasons. Most importantly, perhaps, is that people often measure social standing by their work. Where they work, what jobs they have, how much money they make, and other characteristics of work are important for many human social evaluations. Because this is important, it becomes socialized as something that you should discuss, and thus becomes a common topic of conversation. People then internalize it as something they should talk about, or is interesting to talk about. It’s a self sustaining model built upon the foundations of social worth and evaluation, supported by the emotional needs of humans.
Interestingly you’ll see that in certain circles where social worth is not derived from your work (minorities in which upwards mobility or potential jobs are limited often talk less about work) but from other aspects of your life (talking about children is a favorite for those who have them and artists love to talk about their creative pursuits) that you’ll find conversation drifting towards different topics instead.
I think the best thing you can do, if you find this boring, is to attempt to redirect conversation away from work and towards something you’d rather talk about. People will naturally drift back towards conversation that they find useful, interesting, or have been socialized to do and ultimately you may need to tolerate this or find a group of friends less interested in talking about their career. I’ve generally found that quips which highlight it’s silly to be talking about work away from work (such as when participating in work offsite trips) or highlight how work is just a means to make money and I’m disinterested in talking about capitalism and would rather know the person and what they find interesting tend to work well to divert conversation away from chatting about work.
Not just that, they’re often way less nutritionally useful. Nearly all alternative milks have very low protein content. It would be trivially cheap to add in a small amount of whey protein or use less sugar, but for some reason I haven’t been able to find a brand that does. At most they focus on making it analogous to milk when it comes to frothing for coffee…
That enables companies like this to take advantage of troves of information to set prices while leaving individuals and families struggling to keep up.
Even if individuals had access to this information, it’s being set by a single source. If companies are unwilling to adjust prices to meet demand, and simply set it to what a single source says they should, what leverage do individuals truly have? How many of the individuals will have the time and energy to investigate the source? Once they investigate the source, what actions can they take?
One might make the argument that they can simply not purchase the service, but is this realistic for something like housing? Would you rather be homeless or pay more than you’re comfortable paying? When a company is gigantic enough to survive a significant period of time where they aren’t making profits or has enough holdings that they can float some empty units in order to make more profits int he long term, what levers can affect the way the company operates if they are secure in knowing that no one else will undercut their prices?
At the end of the day what’s lost on these free market fundamentalists is that supply and demand are concepts to describe a **free ** market. If the market is dominated by a single interest, it is by definition not a free market. As you rightly mentioned, most of the time nowadays, larger interests tend to be on the seller/supply side and they have an imbalanced power dynamic with consumers/demand side. This imbalance leads to a market not being a free market, and fundamentalists tend to ignore a nuanced take on power.
Male and female are useful in biology, and therefore in medicine.
As I explained in another comment, in human medicine it is much less useful than knowing what parts a human has and what lab results can tell us about the relevant hormones and other biomarkers of interest. Most people interact with medicine on a personal level, and because of such having more detailed conversations with your doctor(s) will often result in better care.
With that being said, used as a broad term to describe broad effects such as when classifying data at the population level, it can be a useful and quick piece of information to collect. If you’re trying to determine compliance with social determinants of health, it may be faster to collect sex (or gender) than it is to ask people to create a catalog of the important body parts or to ask other broad questions such as “are you disabled” to understand systems better.
It’s an interesting concept, to have a term which is most useful at a certain level of abstraction and less useful the less people you’re referring to with it. We’ve got a decent amount of these in our lexicons and yet I see people drawing false inferences all the time. It’s almost as in if we aren’t having conversations about how broad terms like race, gender, employment status, etc. can be useful when dealing with population level statistics for the purpose of understanding systems, but not particularly useful on an individual basis when trying to determine information about a individual or a small group of them.
I work in healthcare. I’m a data scientist. I get requests all the time where people ask for gender of their patients. Problem is, we don’t capture gender. Or at least, we don’t capture gender for most. We have a field for sex, which is filled in for nearly all patients. Gender is filled in on a separate form which many people are not trained on and thus only present for <5% of our patients.
When I let physicians know that we only have sex available, they inevitably still ask for it. I typically press them as to why- what clinical purpose do you need this for? Their responses vary wildly. Many realize when questioned that they are simply collecting it to collect it - it doesn’t have a real clinical purpose. In some cases, incidences of certain disease states are tied to gender in literature, and knowing that someone is more likely to have a specific disease is something that can be clinically relevant. For these people I provide the information, but I have a short talk with them first. I let them know that the recorded sex often doesn’t tell them what they actually want. There are many individuals with a variety of disorders which can affect what hormones are present in their body, what sex characteristics developed, or how at risk they are for particular disorders. In addition, many trans (and in some cases cis) people may have an inaccurate chart - I have heard plenty of stories of trans men with beards being asked about their prostate by a PCP and trans women asked about concerns related to child birth. While rarer, I have heard the same from some cis people who are androgynous. In most cases a parts inventory is more useful (or in some cases, an understanding of circulating hormones), albeit much like gender, is something we don’t often collect.
Is that all it tells us? Seems a bit of a reach to compare to the usefulness of hot/cold which can inform how/what clothes we should wear to be comfortable or avoid heat stroke or hypothermia, whether an environment can support human life, whether we can get injured from touching an object, what precautions we should be taking when interacting with a hot/cold object, whether a chemical reaction might occur, and many other higher stakes questions than where someone should go to the bathroom.
Exercising judgement is a difficult act, but not one that is black or white. It shouldn’t be painted as something that is or isn’t, either. A slippery slope either existing or not is a false dichotomy trying to shoehorn a complicated situation into an on/off configuration.
Calling the application of social pressure to get cloudflare to stop enabling hate a slippery slope is ignoring that it’s arguably the first instance of something like this to happen, it took an enormous amount of effort for it to happen, while it was not happening the livelihoods of individuals were being harassed, harmed, and destroyed, and it involved a private enterprise making a decision for themselves and is not reflective of how others in the industry will respond.
Of important framing, did we call the workers rights movement a slippery slope? Racial justice? Feminism? I think the more contentious the public perception is of a movement, the more likely people are to call something enabling said movement a slippery slope. However, on the opposite side of things we usually recognize the reduction or removal of human rights or governmental representation universally as a slippery slope when the issue is no longer contentious or is broad enough to apply to all individuals (while nobles may have framed the rise of democracies as a slippery slope away from monarchistic and feudal governmental systems, I doubt the same was said by the majority of individuals who stood to benefit from this paradigm shift). Applying the wording of ‘slippery slope’ to make demons out of issues they simply disagree with seemingly only happens by conservative individuals to protect a worldview that suppresses others.
Perhaps ironically this author could have made their point in a much shorter article. I have no issue paying attention to something long that I feel is of good quality. When it’s a seven minute read because you haven’t bothered to reduce the clutter, or because you make the same statement five times, I’m going to lose interest quickly because I’m not sure you have much to say, but rather you just like to hear yourself talk (or in this case, transcribe it).
Responsibility has to do with control or accountability. Ultimately humans are in control of their actions, so it follows that we are responsible for the outcomes of our actions. How seriously certain people treat this responsibility varies wildly and I would venture that many do not take their responsibility towards other living beings seriously. We also have responsibilities towards objects which are not living as well, and we all suffer from occasionally mistreating the objects which hold value in our lives.
Markdown gives you some control over size, but not by much. I like being able to play with font and sizing to get the look of sections just right. To be honest it’s not that much of a difference and I suspect some of my feelings are just because it’s the first time I’ve tried doing it this way. I think the biggest concern is the 10,000 character limit.
So it’s just a post on our instance, but it was written like a blog article. https://beehaw.org/post/107014
I have no interest in being called a liberal as a slur or by being told who I support when you’ve made no attempt to learn anything about me. This is precisely the kind of behavior we do not tolerate on Beehaw - you have a bone to pick with a certain kind of individual and you’re taking it out on me without understanding the first thing about me. For the record, I think Biden is doing a terrible job and I’m an anarchist.
This isn’t tone policing in the slightest, but I do not have the time to explain or iterate on the nuances relevant here when you’re already painting us with broad strokes. Have a wonderful day.
it will turn into a space full of people seething about communists.
Building a community on being against something inevitably results in that community’s identity being hate for another thing.
To be clear, as an admin of beehaw, we are not explicitly against any ideologies. We do, however, have a clear guiding principle of being nice to each other. Spreading hate and degrading other users are not examples of nice behavior and are not allowed on our platform.
Painting us as ‘against communists’ just because we’ve blocked an instance which does nothing to prevent or discourage specific kinds of speech which are anything but nice, is unfair. We’re not narcissistic enough to believe we’re building the future of community on the internet but we do believe that we are doing something fundamentally different, by centering the kind of behavior we’d like to see and trying to deconstruct the idea of easily subverted rules (too much focus on the letter of law and not enough focus on the spirit of the law).
Fantastic, thank you for sharing this. None of this surprises me as I keep up to date on AI and ethical concerns, but I’m glad it’s receiving more attention.