At the Marketing AI Institute, we read dozens of articles on artificial intelligence every week to uncover the most valuable ones for our subscribers and we curate them for you here. We call it 3 Links in 3 Minutes. Enjoy!
AI Finds Unregistered Voters
According to The New York Times, a multi-state project called the Electronic Registration Information Center (ERIC) has identified 26 million eligible, unregistered voters since 2012 and is playing a part in encouraging participation in this year's midterm elections.
The AI software, which is adept at finding and resolving identities, has also located 10 million registered voters with new addresses, multiple entries on one list, or who are deceased.
Each participating state gives a minimum of voter registration and motor vehicle license data, then the software reviews the data and combines it with supplemental information, like postal change-of-address lists (while also encrypting personal identification information, including Social Security numbers).
With this cleaned up data, participating states can reach out to unregistered voters via direct mail, encouraging them to register.
Machine Learning Deepfakes Could Impact Midterm Elections
Do you think all the political campaign ads you’ve seen on TV, online and on social media are truthful? Rebecca Crootof, Yale law researcher, thinks not—and she’s betting it’s because of artificial intelligence.
In
WIRED
’s article, Crootof explains how she believes AI is going to create an electoral deepfake this midterm: ”A video generated by machine-learning software that shows someone doing or saying something that in fact they did not do or say.”
It wouldn’t be too hard to create, Crootof shares.
Lyrebird
, a voice-cloning tool, is just one example. Lyrebird uses AI and natural language processing to clone voices and create shockingly accurate digital voices, which one could use to say anything. The tool has used created clips of President Trump and Barack Obama “speaking” in an effort to raise awareness of voice-cloning technology.
So far, there has been no public evidence of deepfakes being used to discredit political figures. However, Crootof warns it could become a real issue for the 2020 presidential election.
What is Deep Learning and How is it Different from AI?
The Verge
sat down with Terrence Sejnowski, a computational neuroscientist at the Salk Institute for Biological Studies, to discuss deep learning. What is it? What can and can’t it do? Can we believe the hype?
Sejnowski breaks down deep learning by describing it as sophisticated pattern recognition. Deep learning algorithms are unique because of their architecture—they have very intricate layers that flow through neural networks (modeled after the human brain).
Another unique aspect of deep learning is its ability to solve problems in fields humans never have been able to, like computer vision and translation. This is possible because using different data sets, deep learning can solve multiple problems faster and cheaper than any human.
The biggest hype Sejnowski disproved is the idea that general intelligence and robots are just around the corner. He likened it to the laser pointer. When it was first created, it took up a whole room. Now they’re sold as keychains for five dollars. Self-driving cars, AI and machine learning are here, but it’s going to take some time for incremental advances to make it “more flexible, more safe and more compatible.”