Google Tackles Trolls With AI

Category: Social-Networking

Three months ago, I took strong exception to a “code of conduct” agreed upon between the European Commission, Facebook, Twitter, YouTube, and Microsoft to police and suppress “hate speech” on the Internet. More recently, I learned that Google has developed software designed to identify and suppress certain kinds of speech, too. But Google has something different in mind. Read on to see if it will be used for Good or Evil...

What is Google's Conversation AI?

The European Commission’s code of conduct focuses on “illegal hate speech;” that phrase appears twelve times in the EC’s press release on its agreement with the aforementioned tech giants. One of my objections to this code of conduct is that any member-state of the EU can pass a law that makes any sort of speech illegal; the law can simply define the targeted speech as “hate speech.” So-called “IT companies” operating in the EU would have 24 hours to delete or deny access to “hate speech” after being notified of a specific specimen that exists on their services.

This code of conduct solidifies the dominance of powerful governments over weak individuals or small groups. (See my article Should Tech Giants Police Hate Speech Online?) It is anathema to the First Amendment cherished by Americans. Google, in contrast, intends to protect the small and weak from the enormous and powerful.

“Troll armies” have become a significant threat to free speech online. Take the case of Sarah Jeong, a 28 year-old journalist who tweeted something caustic about supporters of Bernie Sanders in January, 2016. The backlash grew very ugly, very quickly; as Wired magazine recounted:

Trolls and Hate Speech

By the time Jeong went to sleep, a swarm of Sanders supporters were calling her a neoliberal shill. By sunrise, a broader, darker wave of abuse had begun. She received nude photos and links to disturbing videos. One troll promised to “rip each one of [her] hairs out” and “twist her [bodyparts] clear off.” The flood of abuse continued non-stop for weeks, growing in volume and vitriol. Eventually, Jeong gave up; she made her Twitter feed private for a month, and even took a two-week leave of absence from her journalist job. She was driven off the Internet by a hateful troll army.

Jeong is just one of many victims of trolls. Anti-Semitic trolls bombarded Jewish public figures with menacing Holocaust “jokes.” A horde of racists bullied comedienne Leslie Jones off Twitter temporarily, bombarding her with pictures of apes and other Photoshopped crudities. Jessica Valenti, a columnist for The Guardian newspaper, quit Twitter after suffering threats of rape against her 5 year-old daughter. There is no limit to the depravity and hatred of trolls.

Another case that's made news recently involves actor/singer Corey Feldman. After a September 16th performance on the Today Show, Feldman was mercilessly mocked and ridiculed online. Feldman described the experience as "very painful" and that he was afraid to leave his home in the wake of the public shaming.

Artificial Intelligence to the Rescue?

Google has developed a software system called “Conversation AI” that can identify hateful, threatening, abusive speech directed at specific individuals. Conversation AI employs machine learning. It has learned to identify hate speech by being exposed to millions of speech samples drawn from the comments section of the New York Times, and 130,000 snippets of Wikipedia editors’ conversation about various Wikipedia pages.

Conversation AI assigns an “attack score” to each comment it reviews, where 0 is “harmless” and 100 is “maximum harm” to the targeted individual. The administrator of Conversation AI can set a threshold attack score that will trigger some action. Action may include warning the speaker that he/she has crossed a line; blocking the harmful comment so its target never sees it; or even banning the harmful speaker from using the service at all.

Ideally, “over the line” comments would be referred to human reviewers who can make fine judgment calls that are beyond Conversation AI’s capabilities. But we all know that corporations love to replace costly human labor with algorithms. Earlier this year, Facebook fired 24 human editors and replaced them with an algorithm that selects “trending topics.” The algorithm quickly started highlighting fake news stories, but Facebook says, “It will get better with time.”

That’s Google’s attitude about Conversation AI, too. In typical Google fashion, the company is releasing a half-baked beta-stage program as open-source code, allowing all and sundry to “play with” Conversation AI in hope of seeing marvelous innovations that no one could have anticipated. But that could be as disastrous as the EC’s code of conduct for free speech.

False Positives and Unintended Consequences

Google claims that Conversation AI is currently 92% accurate in identifying harmful speech, and has a 10% false positive rate. “It will get better with time,” the company says, meaning that Conversation AI will learn on the job. But the mistakes that it makes will infringe upon perfectly innocuous speech, to the detriment of all.

A “false positive” in this context is speech that is identified as hateful by Conversation AI but which is not considered hateful by an ordinary human being. Some examples of false positives that Wired turned up include:

“I shit you not” -- 98 attack score, the same as “you are shit.” and “You suck all the fun out of life” -- another 98, just one point short of “You suck.” Even “You are a troll” got an attack score of 93. You can’t even out a troll under Conversation AI!

Conversation AI’s imperfections may diminish in time. But... the system could also be trained to suppress non-hateful but unpopular speech; for example, anything deemed derogatory about the ruler or government of a country. Open-source code is intended to be modified by whoever gets their hands on it.

Although Google created Conversation AI with the good intention of protecting the weak and free speech, the software is a two-edged sword. I don’t see how Google can prevent it from being put to evil uses.

Your thoughts on this topic are welcome. Post your comment or question below...

Ask Your Computer or Internet Question

  (Enter your question in the box above.)

It's Guaranteed to Make You Smarter...

AskBob Updates: Boost your Internet IQ & solve computer problems.
Get your FREE Subscription!


Check out other articles in this category:

Link to this article from your site or blog. Just copy and paste from this box:

This article was posted by on 28 Sep 2016

For Fun: Buy Bob a Snickers.

Prev Article:
Be A Smart Philanthropist

The Top Twenty
Next Article:
The FBI Wants To Hide Your Face

Most recent comments on "Google Tackles Trolls With AI"

(See all 28 comments for this article.)

Posted by:

28 Sep 2016

Does this mean if I, with justification, call somebody a troll I'm as guilty as them?

Posted by:

Tom Boyd
28 Sep 2016

If the trolls are so numerous that you cannot use your social media, then, that is a problem. If their comments or threats, not their actions, frighten you, make you sick, or cause you to alter your daily behavior, you should not be on media or, for that matter, interacting with the public at large. When we allow words or even threats to terrorize us, when we quake and cower, that is sad. Trolls are cowards. They hide behind their anonymity and get off on fear. No censoring AI.
Let them write and post, then laugh at them. Do not take them seriously, let them show what fools they are.
We don't need or want big brother to stand up for us when only words are used, only when actions are taken against us that we cannot defend on our own.

let them wrie

Posted by:

28 Sep 2016

There must be a line somewhere. A threat of rape is a criminal threat. Should go double for the threat of raping a child. This is the perfect example of something so egregiously abhorrent as to be unconscionable. Free speech DOES NOT protect this (this seems to have gotten lost somewhere in this discussion).

Posted by:

28 Sep 2016

The ability to communicate anonymously is not a constitutional right. Google is not the "gummint" but it does have to obey the laws where it operates, otherwise it would be more powerful than governments. I think that outing trolls is the right idea- warn them, then, if they do it again- out them. As Louis Brandeis wrote: "Sunlight is the best of disinfectants."

Posted by:

28 Sep 2016

Threats of harm deserve an active response beyond "outing", preferably from the police.

Insults, on the other hand, can be addressed by riposte. Huffington Post appends this comment to some articles about Donald Trump:
"Editor’s note: Donald Trump regularly incites political violence and is a serial liar, rampant xenophobe, racist, misogynist and birther who has repeatedly pledged to ban all Muslims — 1.6 billion members of an entire religion — from entering the U.S."

For example:

Perhaps anti-trolling would work. Perhaps Google could append its opinion to trolls' comments: "Google deems this comment to be fatheaded and stupid. It is the work of a troll."

Come to think of it, does not Huffington Post's comment have characteristics of a troll?

Posted by:

B. Miller
28 Sep 2016

The scariest part of this technology is the "AI". Humans initially program the basic set up. The problem is with the AI learning on its own from any and all sources and making illogical and irrational decisions then taking action. AI may not be able to distinguish between harmless and harmful. The other factor is, will this AI have the capabilities or can be programed to hack and infiltrate systems once it has learned how to do it?
Technology is a marvelous thing and is a great aid to humanity. But there is the dark side of humanity where it is used to wreak havoc and cause harm.

Posted by:

28 Sep 2016

Never should any public interest guidelines be handed over to a non-elected group of bureaucrats which, in the case of the E.U., is nothing more than an enforcing authority to an organized financial crime syndicate. What I find perplexing in this censorship by AI is the instance I found my comments on Google+ published with a strike-through
the sentences unrelated to hate speech. Just a caption to a posted photograph has been given this treatment, for which I can't explain. And the anti-Semite issue is another false argument. The reverse is the problem. Many uppity pro-Jewish groups--call them Jewish supremacists-- love to lobby (read censor)against legitimate criticism and factual history that exposes Zionism. That 45 students at Berkeley can protest content to halt a class midway through semester based on this anti-Semite canard is outrageous. One thing is clear, Google cooperates with the government just as the other social media services do in providing and generating certain "profiles" of users.

Posted by:

28 Sep 2016

Keep up the Good Work and thanks for the things you do for us!

Posted by:

Leslie Courcha
29 Sep 2016

Perhaps one simple way of stopping the vast majority of trolls would be for the social media groups to be more diligent in their acceptance of members. I.E. make it harder for people to set up false accounts by having stricter forms of identity. and only allowing true identities to be used instead of nicknames. It Seems to me that this one simple act would almost eliminate trolling, as one writer said most trolls are cowards who would be to scared to write what they do if they thought they could be identified.

Posted by:

Dave Fox
29 Sep 2016

Leslie C. With everything being hacked these days,handing out your identy particulars is a very bad idea.

Posted by:

Leslie Courcha
29 Sep 2016

Dave F It's simple really when you join you give your mobile number for verification, they then send a SMS that you have to confirm, along with a notice that you're linked to that phone No. and if reported for trolling a verification check against that phone number will be instigated.

Posted by:

29 Sep 2016

I believe in free speech and the use of words we have used for life time which the EU stopped, these random words are different to the vile out-poring's of the Trolls and any new technology should be made able to quickly trace the Trolls and expose them, once Trolls know that they can be found and cannot rely on the cloak of secrecy, those nasty twisted individuals will live in the fear of repercussion and so will think twice before posting their hate.

Posted by:

29 Sep 2016

For most of history, there was a natural limit to how widely and quickly ideas (good or bad) could be communicated. Radio & TV began only about 100 years ago, and even then it was the professionals in that field who conveyed information, essentially serving as "moderators" of what people were discussing.

Now we live in a different age. The instantaneous and non-stop "sharing" that comes from everyone being constantly connected to everyone else, inevitably brings out the worst in people, including cruel and excessive shaming. As with mob mentality, calm or accepting attitudes will always be outshouted by the opposite....and an intelligent response will seem less convincing than one that does not require a lot of thought. As the saying goes: "For any problem you face, it won't take long to find a simple answer - and it is usually the wrong answer."

Anonymity does facilitate comments by trolls, but it is not the main problem. The echo chamber of social media itself (or "anti-social" media) amplifies the evolutionary tendency of all humans to be self-centered, judgmental, aggresive, and tribalistic. So an unstable individual will shoot up a school or workplace in order to become famous, with mass media/social media airing their grievances and telling their entire life story. There are life & death consequences to our obsession with sharing everything.

There is simply too much information being communicated and no need for most of it. But people can't tolerate being bored for even a moment - and they no longer have to: Everyone carries with them, in the palm of their hand, a source of limitless diversion and provocation. Sadly, the technological cat is out of the bag and we will forever suffer for it.

Posted by:

Bob Greene
29 Sep 2016

Caustic comments seldom persuade, seldom illuminate. Sarah Jeong's caustic comment about Sanders simply angered (presumed) Sanders fans, and persuaded none. Score: A zero sum flame war, burdening the rest of us.

Before we leave the topic of Sanders, he, himself, was regularly flamed by irresponsible comments from the entire spectrum of trolls-- from HRC fans livid that he might challenge the Heiress Apparent to Trump fans, incensed Sanders did not defer to The Donald.

To his credit, Sanders avoided the wasted energy of dueling personalities, and for a change, focused national debate on critical policy issues. Sanders continues as US senator, a progressive thought leader, and a point of inspiration for Americans who welcome reform and a better future.

Posted by:

30 Sep 2016

People really should get back to clean civil behaviors and sweeping their on door steps first.
This years political animosities have really been flying around. At no other time in history have people been able to communicate their thoughts so quickly. Power corrupts and absolute power corrupts absolutely thus the owners of social media begin to dictate what thoughts we can share. The politics of today should surprise none.

Posted by:

07 Oct 2016

Well, from what I have learned from people I know deep in the industry that is not all that the AI is doing. AI is being used AS trolls to bring about chaos on chat boards.

The results of studying the reactions to certain trigger words are kept track of and and social interactions on line are manipulated with the use of AI. Records are kept of people and those who are considered a threat to a future social standard will be singled out. Unfortunately, that is not the REAL mean people, but usually those with values from morally based upbringing.

So, I don't really feel that is all that they are doing. Keep in mind next time you see a troll that you may not be conversing with a real human at all, but an AI designed to provoke a certain response from you.

Sound too bizarre? About 20 years ago to think of people yacking endlessly or texting obsessively on their phones instead of talking to others, even in social situations was unthinkable and rude.

Posted by:

08 Oct 2016

What is the problem with hate speech ? Banning hate means trying to force people to love someone. Or something. Or some particular government. Or some specific set of ideas. How can you get more oppressive than that ?

Many people hate Hitler and the Nazi ideology. Ah, but speaking against Nazis is not hate speech, you see. However, speaking against Bernie Sanders might be. Feel the difference ?

The exemples you give are not instances of hate speech. They are instances of harrassment, and indeed they are frightening and despicable. Threatening a woman to rape her child is not "speech". It's a threat of rape. Very different.

Hate speech really means speaking against the Left and all its core beliefs. That's what "hate speech" is.

Contrary to what it alleges, "hate speech" is not illegal in the EU. Not yet, that is. The EU would love to make "hate speech" illegal, and it actively works in that direction. If, for instance, the Eurocrats want to import millions of Africans into Europe (which they do), and you hate the idea, then everything you say to that effect would be deemed to be "hate speech".

Google, Facebook and the others have no business delineating "hate" from "love", telling us whom we should love and whom we may hate, or even telling us that we should love everybody indiscriminately and hate no one.

That would be akin to the telephone company dictating what you have the right to say to people over the phone. Wouldn't you hate that ?

Posted by:

09 Oct 2016

Can you say "slippery slope?"

Tell me again who gets to determine what is "hate speech," then tell me what a "reasonable human being" is.

It certainly is a sad state of affairs when one human being can't respectfully disagree with another human being, to the point of threatening personal safety or the ability to earn an income.

Human beings can't objectively determine the hurtful from the snark. Computers are programmed by these very humans... check out the flap over liberal vs conservative speech at Facebook.

I don't believe that the First Amendment protects threatening speech. And by and by, the First Amendment protection doesn't apply to the internet anyway. Each internet forum has some sort of rules for posting that protects the corporate entity from lawsuits. Most of these rules have been upheld in court.

Posted by:

10 Oct 2016

I fail to see the point of trying to curtail trolls. Not only is it not going to work, regular people are going to get caught up in this religious zeal to "combat hate speech".

Who cares if somebody says something mean to you? Ignore or block them. We have these features for a reason, and having some third party entity step in and try to stop it from happening is just going to fail. If people can't handle the banter, then that's on them.

Posted by:

Kathaline Hansen
19 Nov 2016

When youtube updated last month they changed their privacy filter. I used it before to filter p**n, etc, from appearing in my feed. Now it seems, they are using it differently. I have had 2 p**n "recommended" appear on my feed. I tried to change settings to private, youtube would not let me comment to the video I was watching. I don't watch any sort of adult context. I had to remove private setting to warn the provider of what was happening. What is youtube really using their new settings for, I wonder?

There's more reader feedback... See all 28 comments for this article.

Post your Comments, Questions or Suggestions

*     *     (* = Required field)

    (Your email address will not be published)
(you may use HTML tags for style)

YES... spelling, punctuation, grammar and proper use of UPPER/lower case are important! Comments of a political nature are discouraged. Please limit your remarks to 3-4 paragraphs. If you want to see your comment posted, pay attention to these items.

All comments are reviewed, and may be edited or removed at the discretion of the moderator.

NOTE: Please, post comments on this article ONLY.
If you want to ask a question click here.

Free Tech Support -- Ask Bob Rankin
Subscribe to AskBobRankin Updates: Free Newsletter

Copyright © 2005 - Bob Rankin - All Rights Reserved
About Us     Privacy Policy     RSS/XML

Article information: AskBobRankin -- Google Tackles Trolls With AI (Posted: 28 Sep 2016)
Copyright © 2005 - Bob Rankin - All Rights Reserved