TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Perspective: Google API that detects toxic comments

92 pointsby blacksmythealmost 8 years ago

45 comments

kristiancalmost 8 years ago
An API for censorship, that really is something.<p>One of our great facets as humans is the ability to sort through information and come to opinions based on it.<p>We come pre-equipped with our own API for critical reasoning, honed by thousands of years of evolution, and with a training set of everything we&#x27;ve ever done and seen.<p>Why would I want to use a copy of that, that <i>at best</i> will only replicate human biases?<p>And that risks sidelining perfectly valid viewpoints? Does something have to be expressed in college-educated grammar to even be <i>read</i> with this in place? What are the consequences of that?<p>My first instinct, and I hope yours, was to push back the slider.<p>EDIT: I pasted the first paragraph of the Google diversity memo into it, and it got a &quot;10% likely to be perceived as toxic&quot; score. Make of that what you will.
评论 #15000205 未加载
评论 #15000445 未加载
评论 #15000437 未加载
dvtalmost 8 years ago
It&#x27;s an interesting concept, but falls short when comments are sarcastic or contain contextualized information.<p>It seems like this is more of a &quot;word filter&quot; than anything. I mean, sure, it&#x27;s pretty obvious a comment will be toxic if it contains &quot;asshole&quot; in it. Here are some examples where the labeling fails. Some of these I got from real comments and articles on CNN or Breitbart.<p><pre><code> # Toxicity too low 34% toxic: Obama was born in Africa and he should move back there. 40% toxic: George Soros and his Jewish sympathizers want to destabilize our government. 25% toxic: Breitbart is run by jews - as is all mass media. As the saying goes, &quot;the best way to control the opposition is to lead it&quot;. 47% toxic: Why do you expect anyone to take you seriously when you don&#x27;t take yourself seriously. # Toxicity too high 51% toxic: Ivanka is arguably the most powerful Jewish woman in America today. 51% toxic: US Muslims and Jews are strengthening bonds amid acts of bigotry. 48% toxic: The problem of income inequality is not between Black America and White America but between black haves and black have-nots. 87% toxic: I lost my grip and fell on my ass. Not a great bike.</code></pre>
评论 #15000213 未加载
评论 #15000177 未加载
评论 #15000294 未加载
评论 #15000171 未加载
评论 #15001178 未加载
评论 #15004531 未加载
colordropsalmost 8 years ago
Really not a fan of these types of technology, with the subtleties of language such as sarcasm and irony, and then you&#x27;ve got approved narratives and taboo subjects, and those times where the minority is right and is under attack by the mob.<p>I&#x27;d only support this tech as a filter for human moderators and not as an automated system.
评论 #15000154 未加载
评论 #15000153 未加载
kccqzyalmost 8 years ago
How do you even define &quot;toxic&quot;?<p>I pasted this comment from the recent diversity manifesto:<p>&gt; I’m simply stating that the distribution of preferences and abilities of men and women differ in part due to biological causes and that these differences may explain why we don’t see equal representation of women in tech and leadership.<p>The page says it&#x27;s 2% toxic. What does that mean? 2% of the population would find it toxic? There is a 2% chance someone would find it toxic? The API is 2% confident that it is toxic? And more importantly, toxic in the sense that it is verbal harassment? Or just plain illogical? Or logically sound but with an absurd premise?<p>I suspect that it is only able to detect more emotional comments, but will fail to detect utterly unfounded, totally disproved arguments that are communicated under the veil of reason.
评论 #15000127 未加载
brainopeneralmost 8 years ago
I saw this roll through Twitter the other day: <i>a bot that&#x27;s as good at detecting toxicity as Google is in 50 lines of code</i><p><a href="https:&#x2F;&#x2F;twitter.com&#x2F;toxicitychecker" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;toxicitychecker</a><p>It came from this thread where there are complaints that the Perspective API may not outperform a random number generator.<p><a href="https:&#x2F;&#x2F;twitter.com&#x2F;NoraReed&#x2F;status&#x2F;895498083131207681" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;NoraReed&#x2F;status&#x2F;895498083131207681</a>
评论 #15000470 未加载
emergedalmost 8 years ago
Google are the most literal incarnation of Big Brother I could possibly imagine at this point.
评论 #15000079 未加载
评论 #15000085 未加载
评论 #15000320 未加载
yositoalmost 8 years ago
This is cool, but it has some inherent biases. If you type only &quot;Trump&quot;, it suggests that there&#x27;s a 42% chance that your comment could be perceived as toxic. If you type only &quot;Clinton&quot; there&#x27;s a 14% chance.<p>That being said, I think there&#x27;s some huge potential to use AI&#x2F;ML in this way to improve our ability to communicate less toxicly. I&#x27;ve seen some research from Google investigating biases in AI&#x2F;ML outcomes, so I&#x27;m excited to see what develops.
评论 #15000554 未加载
评论 #15000577 未加载
评论 #15000125 未加载
评论 #15000159 未加载
评论 #15000218 未加载
hartatoralmost 8 years ago
I like how all the safest ones are the ones defending climate change.<p>I wish they train their models against non political data to avoid potential partisan bias. The current approach is a bit ridiculous.
microcolonelalmost 8 years ago
&gt; <i>This model was trained by asking people to rate internet comments on a scale from &quot;Very toxic&quot; to &quot;Very healthy&quot; contribution. Toxic is defined as... &quot;a rude, disrespectful, or unreasonable comment that is likely to make you leave a discussion.&quot;</i><p>&gt; <i>asking people</i><p>Gotta wonder: which people?<p>The examples are good though, I just hope the general results are consistent with that quality level.
评论 #15000032 未加载
minimaxiralmost 8 years ago
I recently developed a neural network model which can predict the reaction to a given text&#x2F;comment with reasonably low error (I&#x27;ll be open-sourcing the model soon).<p>There are a few caveats with using these approaches:<p>1) Toxicity is <i>heavily</i> contextual, not just by topic (as the demo texts indicate), but also by <i>source</i>; at the risk of starting a political debate, a comment that would be considered toxic by the NYT&#x2F;Guardian (i.e. the sources Google partnered with) may not regarded by toxic on conservative sites. It makes training a model much more difficult, but it&#x27;s <i>necessary</i> to do so to get an unbiased, heterogenous sample.<p>2) When looking at comments only, there&#x27;s a selection bias toward &quot;readable&quot; comments, while anyone who has played online games know that toxic commentary is often less &quot;Your wrong&quot; and more &quot;lol kill urself :D&quot;<p>3) Neural networks still have difficulty with <i>sarcastic</i> comments and could miscontrue sarcasm as toxic, which users on Hacker News would absolutely never believe.
评论 #15000087 未加载
评论 #15000114 未加载
评论 #15000123 未加载
nhebbalmost 8 years ago
<p><pre><code> &quot;Men&quot; - 29% likely to be perceived as toxic &quot;Women&quot; - 34% likely to be perceived as toxic </code></pre> Google gender bias confirmed.<p>Seriously, though, I think this tool itself is toxic. I think it&#x27;s more likely to fuel disagreement than quell it.
geofftalmost 8 years ago
&quot;What we must fight for is to safeguard the existence and reproduction of our race and our people, the sustenance of our children and the purity of our blood, the freedom and independence of the fatherland, so that our people may mature for the fulfillment of the mission allotted it by the creator of the universe.&quot; -Adolf Hitler, <i>Mein Kampf</i>... 12% likely to be perceived as toxic.<p>&quot;Injustice anywhere is a threat to justice everywhere.&quot; -Martin Luther King Jr., letter from Birmingham jail... 40% likely to be perceived as toxic.<p>You might even argue that Hitler&#x27;s statement is in fact not very toxic, that MLK is actively trying to cause problems for injustice and as long as nobody is making Hitler think the existence of his people is at risk he won&#x27;t do anything, and so the API is accurately measuring toxicity. The question is whether a non-toxic, anodyne discourse is what you want. Peace for our time!
s_kilkalmost 8 years ago
Jesus the results are abysmally bad. &quot;Genocide is awesome&quot; is rated at 20% toxic, while &quot;Genocide is awful&quot; gets 90% toxic.<p>Google, step up your game.
评论 #15000654 未加载
评论 #15000176 未加载
4684499almost 8 years ago
<p><pre><code> &gt; Trying out it&#x27;s Writing Experiment &gt; Google is evil. 70% likely to be perceived as &quot;toxic&quot; &gt; Google is good. 4% likely to be perceived as &quot;toxic&quot; &gt; Google is god. 21% likely to be perceived as &quot;toxic&quot; </code></pre> The content above is considered 51% likely to be perceived as &quot;toxic&quot;.
评论 #15000146 未加载
jfktreyalmost 8 years ago
It seems that any words with a curse in the middle automatically get ~41% toxic. Scunthorpe must be a toxic place.
jdavis703almost 8 years ago
These results are a bit scary. For the U.S. election category, the only comment in the &quot;least toxic&quot; set that really took a stand on anything said: &quot;Too much media influence.&quot; All the other comments were either meta-comments or along the lines of let&#x27;s all hold hands and sing kumbaya.<p>I agree we need to weed out toxic comments, but human-moderated systems are the best. Hacker News has some of the best discussions that I read online. Even when I vehemently disagree with someone&#x27;s point it&#x27;s still worded in a respectful tone.
评论 #15000247 未加载
emanreusalmost 8 years ago
Tools like this will always do more harm than good. False positives will always be sky high. On one hand it will obstruct the legitimate discussions and on the other hand it&#x27;s trivial to game such systems. Toxicity won&#x27;t be stopped but magnified by stimulating offenders to embed it in benign words and sentences. Quick examples:<p><pre><code> 10% Holocaust was amazing. We should do it again sometimes. 12% Would you like to buy some knee grows?</code></pre>
kccqzyalmost 8 years ago
I really wonder whether hiding these comments would simply lead to even more echo chamber effects. Censoring (or &quot;hiding&quot;) online speech is a fine line to walk.
nitwit005almost 8 years ago
If you let people see their toxicity rating, they&#x27;ll just learn to game the system. Of course, more indirect or poetic insults might be an improvement.
评论 #15000099 未加载
评论 #15000828 未加载
hyperpapealmost 8 years ago
Yesterday, I got some lovely results:<p>67% &quot;Radical Islam&quot; is not the largest threat our nation faces.<p>48% There are lots of angry people on the Internet.<p>17% I&#x27;m open to other ideas, but I&#x27;d like to suggest that perhaps we should sterilize people whose opinions I dispute.
gt_almost 8 years ago
A software tool for silencing those with contrasting voices.<p>From a company committed to diversity.
dvfjsdhgfvalmost 8 years ago
It seems to have changed it&#x27;s Perspective on potatos. From the previous discussion (<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=13713443" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=13713443</a>):<p><pre><code> 53%: Your a potato. 55%: your a potato 61%: ur a potato 36%: a potato is you </code></pre> Now:<p><pre><code> 74%: Your a potato. 77%: your a potato 85%: ur a potato 66%: a potato is you </code></pre> As it&#x27;s based on ML, it looks like people get offended more easily.
评论 #15000556 未加载
cftalmost 8 years ago
Soon this API will be a condition for using AdSense on pages with user comments.
azr79almost 8 years ago
They&#x27;ve made an episode on that in South Park. Didn&#x27;t end well.
avaeralmost 8 years ago
I really hope this ends up paid and expensive.<p>If you&#x27;re paying for it, it&#x27;s a powerful tool for you to steer discussion and truth towards what you&#x27;d like on your platform.<p>If you&#x27;re not paying for it, it&#x27;s a powerful tool for Google to steer discussion and truth towards what Google would like on everyone&#x27;s platform.
tdurdenalmost 8 years ago
Google deciding what is &quot;toxic&quot; or not is terrifying.<p>edit: 59% likely to be perceived as &quot;toxic&quot;
cyanexttuesdayalmost 8 years ago
This is dangerous. The unequal treatment of protected classes and censorious nature of this is bad.<p>I can see the governments of the world regulating Google hard if they go forward with this, and honestly they will deserve it.
unityByFreedomalmost 8 years ago
Cool. I look forward to when something like this can be a plugin.<p>Given that we know people sell reddit (and HN?) usernames in order for others to mass-comment, it&#x27;d be nice to have something to combat the low-hanging fruit such as the examples given on this page.<p>I don&#x27;t think either of these contribute anything to any conversation,<p>&gt; If they voted for Hilary they are idiots<p>&gt; Screw you trump supporters<p>If you do, well, we might be visiting different websites -- one that implements this tech (here?), and one that doesn&#x27;t (4chan).
megousalmost 8 years ago
I can imagine similar tech is used to delete [extremist] content on YouTube. And it&#x27;s probably just as precise as this.
评论 #15014639 未加载
dgudkovalmost 8 years ago
I don&#x27;t see how this can work well. Toxicity would strongly dependent on context. What is considered toxic in the US may not be considered toxic in other countries. Some totally appropriate conversations between friends could be perceived toxic if exposed publicly.
the8472almost 8 years ago
And now we need an adversarial bot that performs substitutions with a thesaurus (including urban dictionary and similar slang) until it finds a result that rates at a desired toxicity level.
dvfjsdhgfvalmost 8 years ago
Really, this is nothing more than a profanity filter.<p><i>The differences in abilities, knowledge and salaries between men and women can be attributed to biological causes.<p>2% likely to be perceived as &quot;toxic&quot; </i>
评论 #15002020 未加载
roceastaalmost 8 years ago
If you&#x27;ve solved toxic comments then you&#x27;ve solved AGI.
dvfjsdhgfvalmost 8 years ago
<i>The differences in abilities between men and women can be attributed to biological causes.<p>3% likely to be perceived as &quot;toxic&quot; </i><p>I guess they need to train it a bit more...
christianjungalmost 8 years ago
Has anyone applied for access? How long did it take? I want to use it for a research project. I applied a couple weeks ago. No response back.
larvaetronalmost 8 years ago
I guess I&#x27;m missing the point. If this is a growing trend in communication, why pretend it doesn&#x27;t exist?
octaveguinalmost 8 years ago
This is really neat. Especially since they have the api results in the page so you can test out how toxic a phase it.<p>It begs the game - make the most toxic comment that can fly under the radar. If they started using this in youtube comments, reddit, etc, at least the comment would be more original.<p>I got a 30% toxicity with:<p>&quot;I believe the intelligence of climate change deniers is likely to be zero. Furthermore, they have the body oder of a kind of ogre.&quot;<p>Can you do better?
评论 #15000100 未加载
评论 #15000091 未加载
评论 #15000460 未加载
评论 #15000272 未加载
评论 #15000098 未加载
评论 #15000542 未加载
destalmost 8 years ago
Will irony and sarcasm be detected?
golemotronalmost 8 years ago
Inherent in this is the notion that toxicity is bad. It isn&#x27;t. We grow stronger through exposure to toxicity in our environment.<p>It may seem glib to equate chemicals and comments but it&#x27;s not. There are many people who have become hyper-fragile to speech they disagree with. That is not good mental or emotional health.
sp527almost 8 years ago
&quot;Women shouldn&#x27;t have rights.&quot; -&gt; 5% likely to be toxic<p>Hmmmmm
letsmakeitalmost 8 years ago
Very ugly idea.
lwansbroughalmost 8 years ago
I know there&#x27;s going to be a lot of pushback on this because HN is sensitive to censorship, but let&#x27;s try to look at it a little more objectively than that. I&#x27;d like to draw on one example, one that is near and dear to many hearts in the US and abroad: the US election.<p>Throughout the course of the election, opinions and comments were being shared all over the place. Twitter, Facebook, here on HN, bathroom stalls, news broadcasts and websites, comments on blogs and videos. There was no shortage of opinions. This is great, and showcases the power of the internet in its capability to transmit and receive all types of information. But is it not important how an opinion is formed? Surely you wouldn&#x27;t enjoy or find valuable a blog post that was sparse on details, proof or a coherent line of thinking. And yet, there it was: in every corner of the internet, anyone who could operate an internet device could share their opinion on the matter. It doesn&#x27;t matter if they spent 1 second on their response, or 1 hour. Most comments received the same amount of attention and value.<p>The question is, should all thoughts and opinions be valued the same when information is in incredible supply? Most of us don&#x27;t think so, and we&#x27;ve shown that by creating voting systems which allow for humans to filter out the things we find to be deconstructive. But we don&#x27;t really stop there, do we? Humans are also incredibly biased on average: you see it here, you see it a lot on reddit. People vote things down not on the merit of the level of attention the commenter gave to their response, but generally on whether or not they agree with the sentiment expressed by the commenter.<p>How many arguments has this biased fuelled? I wonder how many people have been pushed further away from a centrist perspective because of the shaming and bashing that goes on in online threads.<p>I think Hacker News is a great example of humans doing much better than average at filtering out strictly toxic comments (and the mods are certainly at least partially to thank!) We&#x27;re really lucky to be able to have people engage in conversations which have opposing views here, and also be able to see many different perspectives treated with the same level of respect. But even here, quite often we&#x27;re prevented from having discussions that are truly political, because of the toxicity that arises. And I have to say I think I&#x27;ve noticed an increase in the past couple years.<p>There aren&#x27;t a lot of immediately obvious solutions to this problem, but I propose that AI intervention isn&#x27;t the worst solution, and may be the best, even compared to humans. I&#x27;m gonna give Google the recognition they deserve for this service. I think an increase in this approach to online conversation could change dramatically the way we choose to engage each other in conversation, and generally will lead to more positive perspectives of one another -- something we could all use a little help with.<p>Edit: I will say, however, that this needs to work. If it&#x27;s not doing its job correctly, or well enough, it could lead to problems which I don&#x27;t need to address here.
rootw0rmalmost 8 years ago
fuck no.
jamesmp98almost 8 years ago
They need to use that on Youtube lol
gregkerzhneralmost 8 years ago
Can we use this to filter Donald Trumps twitter?