What The Election Taught Me About AI and Media
It was a humbling experience in the last 12+ months. For many of us on the left or right coast, a few of my learning are:
- Online Media are deeply flawed today: Right now we live in a media bubble that is mathematically optimized to show what we wanted to see. Google News, Twitter, Facebook Feed etc have become the primary source of information for many of us — they are great at creating a bubble for each of us to happily live inside.
Media: don’t show me what I wanted to see. Don’t tell me what I wanted to hear. Find ways to give me objective information, please.
2. Data Science has a long way to go and AI might be the rescue: Presidential Election is classification problem. It takes billions of inputs and generates one output. Obviously nobody has figured out a model for it.
This kind of problem is actually a good candidate for AI to solve. I won’t be surprised that none of the existing models used by CNN, Fox, FiveThirtyEight, etc have taken enough advantage of deep learning that has proven to be very effective in the last couple of years (does anyone know this?).
AI should be able to produce deep insight for large scale social/economic issues. Traditionally we relied on deterministic math or statistical models for studying such issues, which have proven to be fairly unreliable. In many ways, deep learning models should be more effective at tackling such issues.
In Life you either win or learn. We learn, evolve and will only become stronger.
ps. Just saw this “An artificial intelligence system that correctly predicted the last 3 elections says Trump will win”
