August 2018    
  Surf and crowdturf  
  Computer scientists Heather Zheng and Ben Zhao used AI to write restaurant reviews.  
  Could bots influence your purchasing decisions? (  
  Artificial intelligence and neural networks--brain-inspired systems that mimic how humans learn--are increasingly commonplace in any research that gleans insights from data. UChicago computer scientists Heather Zheng and Ben Zhao used one such network to expose a cyber vulnerability that threatens many of us: malicious hacking of our dining choices.  
  In 2017 Zheng and Zhao trained a neural network using thousands of Yelp restaurant reviews--teaching the network to write fake reviews that were indistinguishable from real ones. (Can you tell the difference? Take the quiz.) Survey takers rated the reviews as not just believable but also useful, demonstrating that such technology could directly influence consumer decisions.  
  As far as we know, cyber attackers are not yet using AI-powered technology to create fake reviews. Bad actors are still largely employing on-demand crowdturfing systems, where workers are paid to complete malicious tasks.  
  But Zheng and Zhao believe that the threat for AI to automate crowdturfing is real and imminent for companies like Yelp and Amazon, whose customers rely on reviews. And so, using what they learned from creating fake reviews, they are now developing countermeasure algorithms to detect them.  
  Who's writing those Yelp reviews? (  
  Using AI-powered detection to fight AI-powered generation is crucial, Zhao says, because undermining commerce is just the beginning. Artificially created content can shake society's confidence in what is and isn't real. In a 2017 paper (pdf), Zheng and Zhao note that AI can help detect not only fake reviews but also fake news--a problem that has skyrocketed, particularly since the 2016 election cycle. If AI starts generating fake news, it will be especially important to have AI detection at the ready.  
  Check out Zhao's October 11 Chicago Harper Lecture: "The (In)security of Today’s Machine Learning Systems."  
  Six more things about cybersecurity  
The Atlantic explains the Facebook-Cambridge Analytica data scandal in three paragraphs.
Hackers attacked US voting machines at DefCon's Voting Village conference to help secure them from election meddling.
Modern encryption can be traced back to 1882, when a banker was trying to make telegrams private.
UChicago's David Cash uses discrete math and number theory to stop cyber criminals "from doing bad things."
Researchers warn that medical devices like pacemakers are vulnerable to malicious hacking.
Was it data misuse for authorities to search an open-source DNA database to find the Golden State Killer?
  Get to know Incognito  
  How protected is your private browsing window? (  
  Major web browsers offer private browsing windows; Google Chrome calls it Incognito. These private modes keep your searches secret. Right?  
  This past April, UChicago computer scientist Blase Ur and colleagues published a study (pdf) showing that many participants overestimated how much protection these modes offer. Misconceptions persisted even after the participants read a browser's private mode disclosure. For instance, more than half incorrectly thought Incognito browsing kept Google from recording their search history.  
  Think you know Incognito? Take a web privacy quiz.  
  In case you missed it  
Feathered friend or foe: Birds can recognize each other across species.
Win this mug: Forward µChicago and we'll enter you in our monthly drawing.
  Support the Physical Sciences Division.  
  Good things come to those who share! Forward µChicago to a friend and be entered to win a prize. Congratulations to last month's contest winner, Mark Miller, MBA'79.

Sign up to receive µChicago monthly.