This week’s research was pretty intense. I don't have any real experiences with cyberbullying, gaslighting, griefing or any other form of brutality being discussed this week.  I have certainly seen these in action, but haven't personally witnessed anything extreme or very noteworthy from what I can recall.  Certainly nothing like a lot of the stories I've read.  Too many stories about too many young lives ended fare too soon as a result of cyberbullying.  As a parent of three young kids, all at the brink of having their lives inevitably impacted by social media, it is concerning to say the least.  The issues with cyber bullying are well known – it is persistent, pervasive and highly damaging to the psyche of young people not equipped to deal with the issues it presents. 

Before I get into what I spent some time focusing on this week, I want to reiterate what I personally believe to be a fundamental skill for youth this day and age: Critical Thinking.  According to The Critical Thinking Foundation, "Critical thinking is that mode of thinking — about any subject, content, or problem — in which the thinker improves the quality of his or her thinking by skillfully analyzing, assessing, and reconstructing it". [1] Even at their young age it is so easy to see how much social media influences my children’s lives.  Whether it be "that rude guy that was swearing when I was playing Fortnite" or being jealous of that family that "gives their kids all this awesome stuff on Youtube ... like, whatever they want, it's sooooo cool!"  If we don't teach kids how to think, how to really think about the information that is being presented to them, there will be a whole new generation of kids with a real big handicap in life.  Add in mental disorders and how tough it is to be a teenager in general, it's quite a frightening thought ... for me anyway.  One of the technologies that is making this skill so critical now more than ever is Deepfakes.  Deepfake technology is the act of digitally replacing someone’s image or voice in video or audio format.  The technology has come a long way in certain years and is becoming harder and harder to detect while simultaneously more accessible to use.  An excellent example is this well know footage of an AI specialist using an actor to create very life like videos of Tom Cruise.  Or a more recent story about some hackers using a voice clone of a bank director to steal $35 million dollars.  While the technology hasn't become sophisticated enough to be used for nefarious political damage, we are not too far off from that reality. 

In the interim, the technology is more an more being used by people for online harassment and exploitation.  DeepFake pornography is probably chief amongst these issues, with female actors typically the target, but not solely isolated to that demographic.  Earlier this year there was a story of three teenage girls Pennsylvania that were the subject of harassment via deep fake technology.  A woman, who was affiliated with a local cheerleading group, Victory Vipers, sent videos of the cheerleader's to the girls' parents and their local gym depicting them vaping and taking nude videos.  While the motive behind the harassment isn't fully known, the intent was clear: publicly shame and discredit these young woman.  The case was eventually solved, but the turmoil it caused these girls and their family can't be undone.

So what's the solution?  Well ... critical thinking for one!  But obviously that's not enough.  "Deepfake videos are challenging for both face recognition systems and existing detection methods, and the further development of face swapping technology will make it even more so" [2].  However the research in this field is developing as well.  Facebook has recently claimed that it has developed AI technology that can help it detect DeepFakes with very accurate results.  During the "training" for the AI, it learns the generative model used to create the Deepfake and it can then identify Deepfakes and reverse engineer them to also learn what generative model was used to create them (called 'Model Parsing').  Knowing the generative model is key for the technology to further develop and maintain it's sophistication as the deepfake technology continues to advance. 

[1] (2019). Our Concept and Definition of Critical Thinking. www.criticalthinking.org. https://www.criticalthinking.org/pages/our-conception-of-critical-thinking/411

[2] (Pavel Korshunov and Sebastien Marcel). (2018). DeepFakes: a New Threat to Face Recognition? Assessment and Detection. pg 1. https://arxiv.org/pdf/1812.08685.pdf