Published on September 19th, 2018 |
by Chanan Bos
September 19th, 2018 by Chanan Bos
Pravduh about Tesla is an effort that we recently launched here at CleanTechnica after seeing too many misleading, negative headlines about Tesla. The CleanTechnica team spent about a month deliberating and debating how we should approach #Pravduh. There were multiple very important considerations that needed to be decided on and challenges that needed to be overcome.
First and foremost, the system needs to not be biased. So, for a bit of context about what we discussed internally, we tried out multiple systems:
— One system was system scoring: Implications for Tesla (Good, Bad, Neutral, Both). Whether it contains misinformation (Yes, No). Title Connotation, Title Denotation, Text Connotation, Text Denotation (Positive, Negative, Neutral, Both). Whether the news has already been covered by the site before (Yes, No).
The problem was that we did not have the resources to read and score all of the articles. With ~200 articles per week, we realized quickly that was simply beyond what we could do adequately.
— We also considered scoring articles and headlines on a scale of 0 to 5 regarding each of the following: neutrality of language used, quality of sources, logic of conclusions, accurate representation of information, title accuracy. Unfortunately, the scale of work was again too great and being sure to score article uniformly and without bias was sure to turn into a huge nightmare.
Finally, we decided on a system where we rate only the headlines, since most people see only the headlines and make judgements based on them even if they don’t necessarily read the articles.
We rate the titles as either positive, negative, or neutral. While not ideal, this gives a pretty good representation of bias on a news agency level, and possibly also on an author level. We work as a team in a Google sheets spreadsheet and collect the following information: Title, Implication for Tesla, Outlet, Author, Date, URL.
The next part in our Pravduh about Tesla effort is to create a Tesla FUD FAQ page where we would list common misinformation or FUD about Tesla and shortly explain why it is wrong, linking to relevant proof for further reading. In this system, when reading an article, we would give it a score based on how many misinformation/FUD talking points it includes in addition to implications for Tesla and, if possible, tone/slant/bias.
Second of all, the system should eventually be mostly automated.
Here at CleanTechnica, we consider ourselves experts on clean technology, and as an extension, experts on Tesla. Our director and chief editor, Zachary Shahan, was on the Tesla investor call with Elon Musk last month due to our obsession with getting the Tesla story right. We know the score, we know what’s up, and we think we’ve built a good foundation. What we lack, however, are certain resources and skills that we hope you might have, especially programming resources. What we would be looking for in a potential partner is someone who can help us create:
A #Pravduh system that finds all of the news articles on its own.
A system that can do part of the ratings automatically based on a system that compares the article to words or phrases known to indicate misinformation/FUD and certain bias/slant.
With the system automatically putting such information into a database, we could then check the results and create weekly, monthly, and quarterly reports.
Ideally, this system would at a later stage also let people give their own ratings, which could be used to improve our scoring system. Obviously, this system would need to be troll proof, possibly by first giving users a test period and also by letting the weight of their vote increase as they level up — via voting experience, explaining why they vote the way they do, and what others think of their record. It would be nice to also integrate a disqus commenting system for every article we rate, especially for the articles of large news sites that don’t have a commenting system.
The final step — expanding the system to all media #Pravduh.
The final step that we see is the creation of a plugin like disqus that other news websites can integrate that immediately scores the article on slant/bias/misinformation through simple graphs and charts. This plugin would also (either automatically or via the author) link the article to a specific news topic. This way readers could see how the article on this topic compared to the coverage of other news sites (similar to how Google News now gives you all the coverage by different outlets of the same story). Of course, this is a very advanced stage.
If you think you can be of help in some of the ways described above, let us know.
The good news is that people are already finding our weekly reports useful and interesting, including Elon Musk himself. There is much to improve, but like him, we are open minded and interested to see where this leads even using our manual system.
— Elon Musk (@elonmusk) September 17, 2018
Will be interesting to see how this evolves over time
— Elon Musk (@elonmusk) September 17, 2018
Support CleanTechnica’s work by becoming a Member, Supporter, or Ambassador.
Or you can buy a cool t-shirt, cup, baby outfit, bag, or hoodie or make a one-time donation on PayPal.</span>
Share this post if you enjoyed! 🙂