According to information available at the test site, it assesses trustworthiness by analyzing the reputation of the authors who contributed the text, as well as those who edited the text. Each color has a specific meaning; for instance, “[t]ext on white background is trusted text; text on orange background is untrusted text.”
I admit to skepticism both in terms of the process and its effect on readers, so I’ll be eager to hear about the outcome. My concerns are:
* How is reputation assessed since the program depends on it?
* What allowances are there for readers who are color-blind or who have difficulty in distinguishing between certain colors? The program depends on gradations of orange.
* Will color-coding add noise, thereby negating the intended effect?
* Will readers understand the color-coding?