Rand Paul Loses The Plot: Says Gov’t Is Going To Use AI For Censorship
from the that's-not-how-any-of-this-works dept
There are, certainly, all sorts of questions about what is generally referred to today as AI, and what it will do, and if it needs to be regulated, but it would help if we had law makers who had the slightest clue what they were talking about.
Senator Rand Paul is upset about AI and how it might lead to censorship. At first, after hearing this, I thought perhaps he was concerned about how potential problematic regulations of AI could violate the 1st Amendment in silencing speech created by users of AI tools. Or, perhaps, how bad regulations could infringe on the 1st Amendment rights of the companies who make AI tools.
But… no. While those are both legitimate concerns, Paul’s concerns seem… disconnected from reality? It started with a statement during a hearing which already made no sense:
“The government is using your hard-earned tax dollars to surveil and censor your protected speech. Artificial Intelligence is only going to make it easier for the government to do this and harder to detect,” the lawmaker cautioned. “This should not be a partisan issue. We must get to the bottom of how the federal government uses artificial intelligence to violate the privacy and civil liberties of the American people before it’s too late.”
I’m all for protecting the privacy and civil liberties of the American people, but I’m not at all clear what AI has to do with any of this.
Paul then went on a radio show to make things… even dumber.
So if the government decides to use artificial intelligence on Twitter or social media to look for certain phrases that they find to be ‘disinformation’ or the government doesn’t like.
Let’s say, for example, that I say ‘COVID vaccines aren’t really necessary for children that have already had COVID,’ there’s a lot of scientific evidence to back that up, but it’s an opinion and you can have an alternative opinion. But what if those codewords are put into an algorithm? Artificial intelligence then trolls the internet and then the government takes down comments it doesn’t like.
That’s a real First Amendment problem.
So, um, yes, if the government were doing that, that would be a 1st Amendment problem. But it skips a whole lot of steps. First off, what he describes is not AI, it’s filtering systems, and those have existed for years, and the US government has never used such filtering systems to “take down comments it doesn’t like.”
The actual 1st Amendment issue is totally unrelated to the AI here. And, if the government were trolling the internet and taking down comments it didn’t like, there clearly would be a 1st Amendment issue, which is why that doesn’t happen. At most, what we’ve seen is that the government has sent notes to social media suggesting some content violates their policies, or they’ve criticized the companies’ policies. But there has yet to be an example of the government “taking down” content in a way that violates the 1st Amendment.
And… what does AI have to do with any of this? After seeing Paul’s comments, I have no idea what he thinks AI does or how it works, because what he describes… is wholly unrelated to anything related to AI. If the government were censoring content on social media — which, again, it is not — then it’s unclear how AI does literally anything to change that equation. It doesn’t make it any more or less of a 1st Amendment problem. It… doesn’t make any sense at all.
This is like saying “we need to be concerned about mobile phones, because automobiles might run people over, and with mobile phones the government might use them to run people over.” I mean, sure, it would be bad if the government ran people over, but they’re not, and whether or not mobile phones are involved is irrelevant.
Once again, please, I beg of people: stop electing ignorant people.
Filed Under: 1st amendment, ai, censorship, content moderation, rand paul
Comments on “Rand Paul Loses The Plot: Says Gov’t Is Going To Use AI For Censorship”
Texas, Florida, and the Red States: “Write that down! WRITE that down!”
The “AI scary” crowd, whether it’s media, politicians, that dude in charge of OpenAI, or anyone else needs talking points for the scary to make sense to the wider world, and to get rando regulations passed, so they just make shit up.
i’m sure there are real conversations to be had about “AI”, whether what ML does now, or the fantasy AI in peoples’ heads that doesn’t exist, but this ain’t it.
Re:
Okay. Let’s start one here and now with One Simple Question:
What problem does AI solve?
Re: Re:
How to put bodies on the frontline of a war that is about to happen.
Now, technically, you don’t need the current generation of Smarter Markov Chains to force people to die your delusional fantasies, as Putin and Russia have proven.
But forcing people to do so by depriving them of jobs (writers, artists, entertainers, any job that isn’t supporting the war economy) does push it in that direction.
As for “real conversations”…
Has anyone read Dave Karpf outside of his one piece on Techdirtt?
Re: Re:
Keeping the price of GPU’s high, and along with bitcoin mining, keeping them out of the hands of ordinary people.
/S
Re: Re: Re:
nVidia: I am so going to shit out 4060s at affordable prices.
Re: Re: Solutions?
That’s a bit like asking Charles Babbage “But what does your difference engine do, sir? What problem does it solve?”
And Babbage replied “Why, it solves math equations so much more quickly than you or I can, and with far greater accuracy, I assure you. Isn’t that reason enough to assemble such a device?”
Indeed, the difference engine (and its successor, the analytical engine) did have foreseeable uses right from the get-to, unlike today’s AI experiments. However, if we don’t get on the stick and start implementing AI now, how are we going to have a operational Skynet in only six years?
Re: Re: Re:
“and with far greater accuracy”
fwiw, AI makes it sound like it is correct when in fact many times it is not even close.
Re: Re:
How about writer’s block? Feed whatever portion of a book/script you’ve already written into it and tell it to continue. Then – and here’s the important part – use whatever it comes up with as inspiration, don’t use it directly.
Or, you know, if you were to, say, run a weekly D&D campaign, it would probably make an excellent description generator for rooms, caverns, NPCs, you name it.
Re: Re:
That’s already a misleading or leading question. It doesn’t solve one problem specifically, and the term refers to several things. To begin with there’s the whole argument over what’s actually AI and what’s machine learning. Then, it depends on application. You probably already use “AI” without realising it, whether it’s because a logistics system is using it to help optimise routing, a warehouse is using it to improve operations or a website is using it to decide which products to display in the store you visit.
If you’re referring to the latest trend of sticking a text field in front of an AI system in order to create a picture or answer questions for you, then it’s down to the user. I’ve seen people use it for anything from coding templates to writing form letters to creating cover art for self-published books. All of which save time, although the results may still require some editing.
There’s plenty of questions to be answered about it overall, but you have to realise there’s not one question. I remember being asked in the early 90s what the internet was for and being unable to come up with a direct answer, because there was so much it could be used for. This is a similar situation, the problems it solves are the ones people want to solve, which can be anything from automation to solving problems of scale to saving a couple of hours on a project.
Re: Re:
Extend the expertise of experts. It does not replace them,and indeed its results can be dangerous and unreliable when used by non experts. However many fields involve more knowledge than can be held in a single human brain, for example AI trained on symptoms and diagnoses would be a great help to doctors to give them avenues to explore.
Re: Re: Re:
“AI trained on symptoms and diagnoses would be a great help to doctors to give them avenues to explore
OMG, please – no!
Re: Re: Re:2
Couldn’t be worse than people depending on Alex Jones for vaccine advice.
Re: Re: Re:2
To riff on an old saying, an AI is for the guidance of experts and the blind obedience of fools. An AI could save hours of searching references, by referring the doctor to relevant articles, and/or experts. I Don’t expert a doctor to know everything, but I do expect them to be able to use all and any research tools, and medical consultants, to figure out what is wrong.
Re: Re: Re:3
I’d gladly accept that…
Now, if only there was a proper, easily searchable library of medical journals and papers that was not only thoroughly peer-reviewed, but also free to read and attribute…
Re: Re: Re:4
The AI can be trained on those peer reviewed papers, along with adverse drug reaction reports, available diagnostic tests etc. So long as it can provide source references, it will be a better tool than direct searching, especially while the patient is in front of the doctor.
Re: Re: Re:5
Well, first you’d need a proper, easily searchable library of medical journals and papers that was not only thoroughly peer-reviewed, but also free to read and attribute…
Now, where’s that?
Re: Re: Re:2
OMG, please – no!
What could happen? More simpletons causing a run on horse dewormer?
“Once again, please, I beg of people: stop electing ignorant people.”
One way to do that would be to stop publicizing every stupid thing they say. It just makes stupid people more famous.
(Yes, this applies to techdirt doing it also.)
Someone once said (approximately), “I don’t care what they say as long as they spell my name right.”
Re:
Sad but true.
Okay, that part is actually true. Thank you Edward Snowden.
Okay, the current generation of smarter Markov Chains can output disinfo and do so faster than China’s Internet Harassment Army, but a Markov Chain cannot censor.
Rand Paul, are you sure you don’t want to censor your opponents and put them in jail? because I hear Amazon has a device that does just that, BSO has some malware for that, etc…
For an immigrant to become a U.S. citizen, they have to pass a non-trivial test about how our Constitution and how our government works. It would seem appropriate for our Congresspeople to have to pass that same exam AND show a reasonable understanding of the technologies they seek to regulate. I’m dubious that more than a few of the current Congresspeople could pass the citizenship exam, and it’s highly unlikely that most of them could demonstrate any comprehension of computer-related technology and how it’s used.
On that thought, there are a number of pressing issues in this country, ranging from personal privacy to regulatory capture of various government agencies to how Citizens United has allowed a flood of super PACs to seriously impact our elections to how much of our money is being stolen by the ultrawealthy. And that’s not to mention some basic personal freedom concerns, such as whether or not we have control over our own bodies. Since Congress doesn’t seem inclined to address any real issues and prefers to chase stupidities, e.g., banning TikTok, I suggest it’s time we stopped paying them! If I have an employee who doesn’t do the job, they don’t work here any more. If I have a contractor who doesn’t perform as contracted, I sue them. I sure wish we could sue every member of Congress for failure to perform their duties!
sad
Didnt he get a concussion that time the neighbor attacked him over his yard leaves? Maybe it was worse than they thought.
or maybe...
He is using a chatbot to write his speeches?
Look, guys. Have you got something for me, or do you not? The War on Terror has worn down, people are tired of it. War on Drugs? Ugh. Don’t get me started. If I’m going to get reelected, I’m going to need to be riding the Next Big Moral Panic. …Yes, meaning we’re going to be manufacturing it. Of course we are. Now, what ya got?
— Rand Paul, reelection strategy session. Probably.
What are you doing, Rand?
Rand Paul is upset about intelligence of any kind, because it goes against his Libertarian principles. Ignorance is bliss, and Libertarians are very blissful people…
AI will lead to even more groupthink than we have now.
the US government has never used such filtering systems to “take down comments it doesn’t like.”
?
He's just representing (some of) the people that elected him.
Ron Paul’s son came to the state I live in because he thought they’d elect him. While his disaffection with McConnell (another guy who came to Kentucky to gain elected office) is genuine, it was McConnell that taught states how to apply that newfangled GIS to demographic analysis. Gerrymandering crammed and cut the then Democratic majority of Kentucky into polarized districts so that many parts of the Commonwealth became safe Republican districts. Paul, to some degree, is a product of McConnell’s work. A few decades of legislators that didn’t need to attend to the will of their constituents has fostered the polarization we now witness in the likes of Mr. Paul. It feeds the trope that government is bad for people. Understanding the mechanics of Kentucky’s governance throws some light on the common desire to elect better public servants. Wishing isn’t going to deliver much. Expecting Paul or McConnell to repair their political malignancies is wishful thinking too. Given the national damage and the absurdity of thinking the folks elected in Kentucky will dismantle the scheme that elected them, I can only suggest that it needs a national scale solution. Democracy as designed in the US needs the elected to feel unsafe about their jobs so they work to find a practical compromise. The status quo, post McConnell, is a systematic polarization that weakens us all.
He's right you know
If they take down 230, everyone will use AI to censor.
Re:
If 230 is reversed, will websites be liable for postings of AI?
Wait .. can AI agree to the TOS? If not, they can not use the resource.
Of all the ‘hypotheticals’ to use that he went with that one says so very much about him and his target audience…
Can I have a paragraph written in the tone of a paranoid conservative ranting about the dangers of AI
ChatGPT
Folks, let me tell you, we’re standing at the precipice of a technological abyss, and it’s time to wake up! This so-called “artificial intelligence” is nothing more than a diabolical creation that threatens the very fabric of our existence. It’s like a Frankenstein’s monster, lurking in the shadows, waiting to devour our freedoms and control our lives. These tech elites and their AI cronies are hell-bent on turning us into mindless drones, manipulated by algorithms and devoid of independent thought. They claim it’s all about convenience and efficiency, but mark my words, it’s a Trojan horse of tyranny! They’ll invade our privacy, monitor our every move, and even predict our actions before we’ve even thought of them. Our thoughts, our preferences, our secrets—all will be harvested by these soulless machines. We’ll become slaves to the whims of cold, heartless algorithms, losing our individuality and autonomy in the process. Don’t be fooled by their promises of progress; this is the slippery slope to a dystopian nightmare! It’s time to resist, my friends, and safeguard the future of humanity before it’s too late. We must stand against this AI juggernaut and reclaim our sovereignty, our privacy, and our human spirit!
At least we know it can mimic GOP word salad with no problems.
It’s Rand Paul. He never had the plot to lose.
And he’s flooding the endzone with shit. It’s all he ever does. The best thing to do is to ignore the ignorant little bastard.
This comment has been flagged by the community. Click here to show it.
Mike Masnick lies again
“the US government has never used such filtering systems to “take down comments it doesn’t like.”
Not true, the FCC demanded that phone companies ban robocalls, robotexts.
In contrast a twitter employee caught by Project Veritas, admitted that they used machine learning to label some people on the platforms as bots, based on their political viewpoint.
Moreover there are members of congress, who think that social media needs more automated viewpoint censorship, when they communicated with the twitter corporation, in addition to speeches and debates in congress.
Re:
“the US government has never used such filtering systems to “take down comments it doesn’t like.
Not true, the FCC demanded that phone companies ban robocalls, robotexts.”
“In contrast a twitter employee caught by Project Veritas”
“Moreover there are members of congress ..”
Re:
The second you used Project Veritas, you instantly get handed a red card. That shit doesn’t fly anywhere except with inbreeds.
Re:
“Not true, the FCC demanded that phone companies ban robocalls, robotexts.”
That’s not so much “comments it doesn’t like”, but unsolicited harassment. The correct online comparison would be spam, and those rules should be applied everywhere, no matter the content of those calls and texts.
“caught by Project Veritas”
If that name’s involved, you can always assume correctly that something’s been manipulated and misleadingly presented. If you get your news from places that present it their reports as unvarnished truth, you should consider expanding your news diet.
“social media needs more automated viewpoint censorship”
No, they believe that targeted bullying and abuse based on race, gender, sexuality and other things should be something to be tackled, along with deliberate misinformation, and in the real world there’s no way to deal with that on platforms with hundreds of millions, if not billions, of users, is with some form of automation. I wouldn’t want the US congress to dictate the tech, but it’s needed at that scale.
You should probably arm yourself with facts before accusing others of lying.
Re:
Not true, the FCC demanded that phone companies ban robocalls, robotexts.
Ya know, for all of the supposed “censorship” going on, why is it that I still see all of this verbal diarrhea from you goddamn people over and over again? It’s almost like you’re not being censored enough, with all this complaining that’s getting through.
To the people doing the censoring – can you ramp it up a bit and shut these fucking complainers up? I’m sick of their bullshit about being censored not being adequately censored.
Re:
Oh, it’s the revenge porner again.
Try running your bullshit, clearly treasonous arguments against the Kiwifarms legal peanut gallery, and DON’T COME BACK until you can get your legal nonsense correct against THEM.
Re:
First, robocalls and robotexts are not comments. Second, the issue isn’t whether or not the government likes them; it’s about the fact that they’re auto-generated. That’s not content-specific.
PV is a known liar who distorts what they say. I would trust Fox News over them.
What certain congresspeople want is irrelevant to what has actually occurred.
AI doesn't cause censorship, but can make it easier.
The AI language models aren’t used only for generating text, they’re also used for classifying it. If you look at Internet fora under heavy censorship such as in China, you can sometimes see people using indirect, obfuscated language tricks to avoid the censors. It takes a huge amount of human labor to find them all and thoroughly shut down the conversations you don’t want happening. A sufficiently advanced AI language classifier can do it all automatically.
Of course it will come with a lot of hidden biases and false positives, but that may be an acceptable cost — as it is for proponents of #chatControl in Europe for example.
Re:
It’s a tool. It has many positive uses. It has many negative ones. But, it exists now, so we should deal with the realities of how it’s uses and the questions raised.
Re: Re:
Nuclear weapons are tools that exist, doesn’t mean everyone should have one.
This comment has been flagged by the community. Click here to show it.
Rand Paul is always going to be more right than you
He is simply pointing out that various government agencies are engaged in efforts to suppress speech on Social Media (a fact you consistently deny and lie about).
He’s simply pointing out that the FBI (and others) should be denied those tools.
Rand Paul is a pretty smart man. (I’m openly a fan) You on the otherhand, are an idiot not even understanding what your betters are talking about.
Techdirt: Look at all these various technologies the US government has used to violate people’s civil rights!
Also Techdirt: The US government? Use AI to violate civil rights? That’s disconnected from reality!
Re:
Look at that man over there! Some say he’s made of straw, but I say he’s coming for your children!
Re: Re:
okay booster