Quicksilver is a machine-learning tool from AI startup Primer: it used 30,000 Wikipedia entries to create a model that allowed it to identify the characteristics that make a scientist noteworthy enough for encyclopedic inclusion; then it mined the academic search-engine Semantic Scholar to identify the 200,000 scholars in a variety of fields; now it is systematically composing draft Wikipedia entries for scholars on its list who are missing from the encyclopedia.
Read the rest
The always entertaining and inspiring YouTuber Code Bullet decided to build a 2048-playing AI from scratch. It's really interesting to see him also build 2048 as part of the process. Read the rest
Can a "griefbot" help you mourn?
In recent years a few computer scientists have created chatbots of deceased loved ones, by training AIs on transcripts of the deceased's online utterances. There's the case of Roman Mazurenko, a Russian man whose friends created a chatbot based on his texts; there's Muhammad Aurangzeb Ahmad, who similarly constructed a bot of his father so that his children would have some sense of what it was like to talk to him.
It's a form of mourning and remembrance that's quintessentially modern, and raises interesting questions about what the shape of grief will look like in the years to come. These experiments in griefbots thus far have all been bespoke, but I doubt it'll be long before we see one-click bot-creation – where you feed a service the various screen names and accounts of the deceased, and it's all autoscraped and assembled quickly into something you can chat with.
But what's the emotional impact of talking to a chatbot version of someone when you know it's just a bot? My friend Evan Selinger is a philosopher who writes frequently and thoughtfully on the moral implications of tech, and in a recent essay he suggests an intriguing parallel: The "empty chair" technique ...
Read the rest
The empty chair technique that I’m referring to was popularized by Friedrich Perls (more widely known as Fritz Perls), a founder of Gestalt therapy. The basic setup looks like this: Two chairs are placed near each other; a psychotherapy patient sits in one chair and talks to the other, unoccupied chair.
The DoNotPay bot (previously) is a versatile consumer advocacy chatbot created by UK-born Stanford computer science undergrad Joshua Browder, with its origins in a bot to beat malformed and improper traffic tickets, helping its users step through the process of finding ways to invalidate the tickets and saving its users millions in the process.
Read the rest
On January 17, 2017, Girl 4 Trump USA joined Twitter. She was silent for a week, but on January 24, she suddenly got busy, posting an average of 1,289 tweets a day, many of which were in support of U.S. President Donald Trump. By the time Twitter figured out that Girl 4 Trump USA was a bot, “she” had tweeted 34,800 times. Twitter deleted the account, along with a large number of other Twitter bots with “MAGA,” “deplorable,” and “trump” in the handle and avatar images of young women in bikinis or halter tops, all posting the same headlines from sources like the Kremlin broadcaster RT. But Twitter can’t stop the flood of bots on its platform, and the botmakers are getting smarter about escaping detection.
What’s going on? That’s what Sam Woolley is finding out. Woolley, who recently joined Institute for the Future as a Research Director, was the Director of Research at the Computational Propaganda Project at Oxford University. In this episode of For Future Reference, we asked Sam to share highlights of his research showing how political botnets — what he calls computational propaganda — are being used to influence public opinion.
Listen to the podcast interview with Sam Woolley here. Subscribe to the IFTF podcast on iTunes | RSS | Download MP3 Read the rest
I woke up this morning thinking the world needed a Prince Vultan tweet-bot.
You know he'll say it sooner or later...
Jack Burton is still talking to whomever is listening, and Sherman T. Botter contiues to share Potterisms. Read the rest
Spamnesty is a simple service: forward your spam to it and it will engage the spammer in pointless chatbot email chains, wasting their time.
If you get a spam email, simply forward it to firstname.lastname@example.org, and Spamnesty will strip your email address, pretend it's a real person and reply to the email. Just remember to strip out any personal information from the body of the email, as it will be used so the reply looks more legitimate. That way, the spammer will start talking to a bot, and hopefully waste some time there instead of spending it on a real victim. Meanwhile, Spamnesty will send you an email with a link to the conversation, so you can watch it unfold live!
The conversations are indeed posted live, and some are quite funny. It's fascinating how obvious it is when a spammer switches from their own bot to giving a human response, and satisfying to see them fooled.
Have you met Lenny? Read the rest
After a series of scandals and rumors, the Hall of Presidents at Walt Disney World has reopened with its Donald Trump robot, which "features the latest advances in technology that enable smoother and more lifelike movements" and "personally recorded remarks exclusively for The Hall of Presidents."
Read the rest
Mark writes, "Nothing marks the holidays like the predictability of a formulaic chestnut featuring '90s stars, magical religious holidays, SFW romance, good hair, and reliable stable camera work. For all those who need a bit more than TV can deliver, my kids and I created the Hallmark Holiday Movie Bot, which generates one feel-good Chrismukkah hit after another for your seasonal celebration!"
Read the rest
The Botometer is a simple single-serving website that reports whether any given Twitter account talks like a bot. It seems quite accurate, tracking not just the content but "sentiment" and its networking characteristics.
My account, @beschizza, has a "green" score of 38%, so I have passed my Twitter Voigt-Kampff test. But @boingboing scores 53%, perhaps reflecting its mix of human chatter and automated links to posts. (Trump also scores 53%, oddly enough.)
Read the rest
How should I interpret a bot score?
Roughly speaking, one can interpret a bot score as a likelihood that the user is a bot. As such, bot scores closer to the extreme values of 0% and 100% are more confident assertions of the account's bot-ness.
It's tempting to set some arbitrary threshold score and consider everything above that number a bot and everything below a human, but this is probably not the best way to think about it. Many accounts score in the "yellow" range of 40-60%. A middle-of-the-road score like this is a signal that our classifier is uncertain about the classification.
In The Guardian, my Institute for the Future colleagues Marina Gorbis and Sam Woolley write about social bots as a threat to democracy:
Social media platforms may be able to track and report on political advertisements from foreign entities, but will they divulge information on pervasive and personalized advertising from their domestic political clients?
This is a pressing question, because social bots are likely to continue to grow in sophistication. At a recent roundtable on the Future of AI and Democracy, several technology experts forecast that bots will become even more persuasive, more emotional and more personalized.
They will be able to not just spread information, but to truly converse and persuade their human interlocutors in order to even more effectively push the latter’s emotional buttons.
Bring together advances in neuroscience, the ability to analyze massive amounts of behavioral data and the proliferation of sensors and connectivity and you have a powerful recipe for affecting society though computational means.
"Social media bots threaten democracy. But we are not helpless" (The Guardian) Read the rest
Up until now, Russian and other foreign bots have been chummy with Trump and the GOP on social media, with "nearly half of the president’s followers appearing to be fake or spam accounts," according to Newsweek.
But over the last 48 hours, these bots have begun to change their tune. According to Newsweek:
Russian-linked bots and trolls have caused a surge in use of the hashtag #ResignPaulRyan on Twitter over the last 48 hours, just as the Republican speaker of the House was returning to his hometown of Wisconsin for a month-long respite from Washington, D.C.
A monitoring dashboard established by the Alliance [German Marshall Fund’s Alliance For Securing Democracy] noted the uptick Monday morning. It coincided with surges in the use of other hashtags by Russian bots, including #TrumpTV, #Magnitsky, #Fake and #ConfessYourUnpopularOpinion.
It hasn’t been a common occurrence for the Alliance’s dashboard to pick up on Russian bot activity targeting members of the GOP since the site was first launched last week by former FBI special agent Clint Watts. But Ryan wasn’t the only member of Trump’s party to face countless bots demanding his removal. A campaign calling on the president to fire National Security Advisor H.R. McMaster received widespread support from bots and trolls over the last several days using the hashtag #FireMcMaster, eventually getting picked up by some right-wing fake news sites that seem to have the president’s full attention, including Breitbart.
Nice time to go on a 17-day vacation.
Image: DonkeyHotey Read the rest
Twitter's indecisive approach to dealing with trolls, harassment and general abuse—suspected by the paranoid as a symptom its need for growth and reach—confounds users to this day. But the blind eye enables more interesting phenomena, too, such as bot armies pushing fringe stories into the trending tags list.
MicroChip, who operates behind a VPN (a special secure network that obscures his location), is an object of fascination and fear, even among some of his political and ideological fellow travelers, who hope not to end up on the wrong side of one of his Twitter campaigns. One conservative observer of the alt-right, who spoke to BuzzFeed on the condition that his name not be used, claimed he once hired private investigators to trace him. ... MicroChip said the truth, both about his identity and the method he developed for spreading pro-Trump messages on Twitter, is far more prosaic. Though he would not divulge his real name or corroborate his claim, MicroChip said that he is a freelance mobile software developer in his early thirties and lives in Utah. In a conversation over the gaming chat platform Discord, MicroChip, who speaks unaccented, idiomatic American English, said that he guards his identity so closely for two reasons: first, because he fears losing contract work due to his beliefs, and second, because of what he calls an “uninformed” discourse in the media and Washington around Russian influence and botting.
The alt-right botmaster describes himself a "staunch liberal" who was "redpilled" by Islamic terrorism, then figured out how to automate Twitter trends. Read the rest
Biomimicry continues to make amazing strides. Festo just released footage of their OctopusGripper being put through the paces. Read the rest
In Even good bots fight, a paper written by Oxford Internet Institute researchers and published in PLOS One, the authors survey the edits and reverts made by Wikipedia's diverse community of bots, uncovering some curious corners where bots -- rate-limited by Wikipedia's rules for bots -- slowly and remorseless follow one another around, reverting each other. Read the rest
Zuck That says, "Have you ever been on the Internet when you came across a checkbox that says “I’m not a robot?” In this video, I explain how those checkboxes (No CAPTCHA reCAPTCHAs) work as well as why they exist in the first place."
Read the rest
I mention CAPTCHA farms briefly, but the idea behind them is pretty straightforward. If a company wants to create an automatic computer program to buy 1,000 tickets to an event or make 1,000 email accounts, they can make a script that fills out the form one at a time, and when the program gets to a CAPTCHA, it will send a picture of it to a CAPTCHA farm where a low-wage worker will solve it and send the answer back to the computer program so that it can be used to finish filling out the form.