Influencers of Moments
I’ve done some reflecting after launching the What’s Good Project and writing about the Theory of Moments. What’s Good leveraged Google Machine Learning and Natural Language Processing API to analyze and score the sentiment of news articles on a -1 to 1 scale. I used machine learning to create a new service tier for Rossella that enables customers to filter their scheduled tweets by tone: negative, neutral, positive or very positive. AI and Machine Learning are powerful tool in analyzing social media content to identify bias, bots, and fake news.
What can we do to create smarter and more accurate bias and fake news detection tools?
- Analyze the Distribution of Moments
- Entity-Sentiment Analysis w/ Entity Clusters
- Link Source Analysis
- Social Network Analysis
I wrote about (1) in the Theory of Moments. I am working on (2) for a #TrollPurge project launching today. Method 2 uses Google Machine Learning text analysis to identify and cluster related entities and then calculate a cumulative score, magnitude, and salience values. This will tell us to what degree a social media account or news article is biased against the “Democratic Party”, “Barack Obama”, or the “European Union”. I would have included some right-leaning US or global entities but that ignores the reality of what happened since the 2016 election.
The quality of social media content links (3) is another factor in determining the status of an account. An account that shares links from credible sources including Pulitzer Prize-winning media organizations, websites with high Alexa scores, and popular, credible blogs.
The Bad News
The wunderkinds of Silicon Valley are too busy counting money to care about things like truth and democracy. Taking a wider view, Twitter, Facebook, and Google operate in many countries around the world that restrict freedom of information online including Turkey and China. It’s no surprise that Silicon Valley has turned a blind eye to Russian interference and the proliferation of bots in everyday life on social media. Yet, they really, really don’t give a damn about managing trolls and harassers, e.g., GamerGate, Neo -Nazis Get Stars, etc.
The Good News
Entrepreneurs and innovators are stepping up to the plate. TwitterAudit enables you to quickly determine the percentage of real vs. fake accounts that follow you. Botcheck.me analyzes Twitter accounts and identifies bots spreading propaganda.
We The People need to strike back. Perception-altering fake news is a pipeline problem. We must identify, mute or block, and report bots and trolls. We also need to outshine their darkness by sharing more news from trusted sources. Together, we can eliminate the threat posed by bots, trolls, and fake news!!
