AMA stands for Ask Me Anything. It’s a crowd-sourced interview through the popular content sharing platform, Reddit. Assistant Professor of Psychology and Global Urban Studies at Michigan State University, Zachary Neal, and former Michigan State University artificial intelligence researcher (currently at Pennsylvania State University), Randy Olson, came together to answer questions about data science, artificial intelligence, and how data is visualized and shared. Here are the my top five takeaways from their interview.
Urban planning and data science go hand in hand from building a community to building an economy.
Professor Neal: There’s a lot of potential overlap between urban planning and network science, but it’s been under the radar until fairly recently. I wrote my book “The Connected City” for a general audience to try and highlight some of the potential. At the micro-level, understanding networks can shed light on how neighborhood communities form or dissolve. At the meso-level, the structure of street networks shapes how different cities (or different parts of the city) are experienced by residents and visitors. And at the macro-level, which has received the most attention, the structure of global transportation and finance networks have important implications for sustainability under the threat of economic or epidemic outbreaks.
Cities use data science to find innovative solutions to crime, but can be hindered by politics.
Professor Neal: Cities are already using data to address issues like crime and community-building, though some may be more effective than others. In some of my other work, I’ve used simulation models to explore the conditions under which strong urban communities form (see: https://www.msu.edu/~zpneal/publications/neal-bigsmall.pdf). A key challenge is that a data-based solution to an urban problem must also be politically tenable to have a chance at being implemented.
We are all wrong when we try to predict emerging technologies.
Randy Olson: I generally avoid trying to predict the progress of emerging technologies because we’re all inevitably wrong, but I’d love to see an affordable robotic butler within 10 years. I wrote about that in my graduate school application and think it should be feasible within the next decade.
No, Skynet is not going to happen any time soon. But, we should be concerned about artificial intelligence in weapons.
Randy Olson: At first, I was very critical of Hawking and Musk speaking out against AI and projecting the stereotypical “AI will conquer us all!” fear that the media loves to write about. But after reading the open letter that they were a part of, I understand their position and actually agree with them.
To be clear: There is no imminent danger of a Skynet-like AI taking over the world anytime soon. If anything so ambitious were being undertaken, we’d know about it, and the researchers wouldn’t do something silly like connect it to our missile defense systems.
When it comes to data science, sometimes we try too hard to make it look nice and that hurts the science.
Professor Neal: I am excited to see that data visualization is growing as rapidly as it is, particularly because it helps tell a better and clearer story with data. But, when it comes to looking for the best data, I am concerned that too often the goal is for the largest or newest or cleanest data, without a particular concern with whether the visualizer really understands the context of the data.