Beyond a Moment of Silence: How STEM Can Prevent Gun Violence | Divya Nori | TEDxAlpharettaWomen
URL: https://www.youtube.com/watch?v=gEyZaz9Aq3k Video ID: gEyZaz9Aq3k ============================================================ [Music] i remember hearing about a school shooting for the first time when i was nine twenty children and six staff members died at sandy hook elementary school in 2012. my mom told me that seven-year-olds had died while they were at school and although that felt scary as a fourth grader i didn't understand what that meant i didn't understand that these kids went to school on the bus just like any other day but they never came back but six years later i did understand in 2018 a gunman opened fire in parkland florida he took innocent lives at marjory stoneman douglas high school my school participated in a walkout we held 17 minutes of silence to honor the 17 lives lost and over the next month three thousand schools participated in walkouts and protests that's over two million students but as news headlines began to change attention on the tragedy and gun violence prevention began to slip into the background parkland is just the next state over from where i live but inside the bubble of milton georgia gun violence didn't seem like a pressing concern we were close by distance but far by experience just three months later another devastating shooting occurred at santa fe high school i experienced a wake-up call an infinite number of silent moments can't save lives gun violence is an ongoing concern and we can't let the presence or absence of a national headline affect the extent of our activism news articles describing the tragedy said that the santa fe community was unsurprised the perpetrator posted a facebook photo just days before the shooting wearing a t-shirt that said born to kill previous shooters show similar signs for example the 2019 dayton shooter liked and retweeted several tweets about a previous shooting in el paso texas the parkland shooter made a disturbing instagram post indicating that he planned to carry out the shooting even the sandy hook shooters obsession with violent video games could have been a warning sign but how far is too far is tweeting about a fascination with guns concerning how about posting a picture with a disturbing t-shirt looking at these qualitative characteristics and trying to draw serious conclusions from them can be misleading and that is where technology can help technology's greatest asset is analyzing large amounts of data and finding trends in a short amount of time so using technology we could quantify whether an individual is high risk for violent behavior previous studies show some important results when mass shooting data from the past 55 years is analyzed we can identify a few patterns first 75 percent of shooters showed unusual interest in past shootings eighty percent of shooters showed some signs of suicidal ideation we need to think about these results cautiously the relationship between these patterns and gun violence isn't causal there's something more complex going on here and technology can help us detangle and quantify that i imagined this type of algorithm we look at a message or a phrase that could have been posted on social media and we break it up into words each word is assigned a number based on how strongly it indicates one of those two concerning patterns glorification of violence or suicidal ideation that number is called a word's coefficient when a word has a positive coefficient that means it's indicative or predictive of one of those patterns for example in this phrase i was born to kill the word kill has a coefficient of positive 0.8 if a word indicates the opposite of those patterns such as happy or life then it would have a negative coefficient but how do we actually come up with these numbers the answer is with vast amounts of data i collected over 1.6 million tweets using computational techniques and this was the input for three artificial intelligence models built in python the output of the models was words and coefficients let's go back to the sample message using the coefficients from the model we can calculate a final score it's not a simple addition expression as shown here but we can think about it that way that final score will always be between zero and one where higher scores indicate concern so at this point you might be thinking there are a few problems with this approach for example what about a misleading word in this phrase these guns make me happy the model might get confused because the word happy typically isn't concerning one way that we tackle this problem is by assigning coefficients to groups of words not just words the word happy alone has a coefficient of negative 0.8 but that whole phrase guns make me happy has a coefficient of positive 0.95 so that whole phrase actually has a highly positive final score and it's therefore deemed concerning another problem is teenagers are often sarcastic especially on social media in this phrase math is so hard i want to die lol it may not indicate genuine concern unless it's backed up by other messages one way that we avoid this problem is by training the artificial intelligence models with tweets those tweets are curated from a millennial and gen z audience so it already contains a lot of sarcasm the model is able to learn from these sarcastic patterns and apply its conclusions to new messages another solution is a score threshold the score threshold is essentially a value and if a message's final score goes above that value then it's deemed concerning although this phrase might have a positive score due to the word die its score likely wouldn't surpass the score threshold so at this point we have a model that can take in a message and give us a score how do we use that to create a real tool for users i created a mobile app the app monitors a teen's outgoing text messages and scores each one using the model if the score surpasses that threshold then a parent will receive an alert here's a quick demo of the app here on the left i have a sample child's phone and on the right i have a sample parent's phone on the child's phone we can go into the regular sms app and first we can send a message that isn't concerning for example want to grab lunch as you can see over on the parents phone there isn't an alert sent next we can send a message that contains sarcasm once it's sent the model will score the message and if that score surpasses the threshold the parent will receive an alert but there isn't one lastly we can send a message that is more concerning and almost instantaneously the parent receives an alert we can see that the alert says that the average score has gone above the threshold one concern with this type of app is the user's safety with personal information being monitored these ethical challenges need to be tackled while there's a lot more work that needs to be done in this area the app does have a measure in place when a teen's outgoing text message is monitored the scoring process takes place directly on the phone the content of the message will never leave the phone only a numerical score in this way the user's personal information is being protected there's a lot more work that we need to do before this app can become a widely available solution but this proof of concept shows that science technology engineering and math or stem can have a positive impact on mass shooting prevention but before the stem perspective can be fully leveraged we may need to shift our mindset as a community gun violence is often perceived as an issue that's solely tackled through political activism and although this traditional approach is important stem offers an untapped perspective that we can use to tackle this issue and many other social issues together we can become more aware of these stem solutions and even strive towards them as a young person living in times when mass shootings are devastatingly common i believe we as a society are obligated to create a safer world for the generations after us and with each of our unique strengths we are up for the challenge every one of us has a role to play to make sure that never again isn't just a trending hashtag we can all contribute to making sure that mass shootings truly occur never again as stem helps pave the road towards a world that is free of gun violence we can all affect change beyond a moment of silence thank you [Music] you