Thread by @breadandnet

⚡️What's on now?

🔬A Citizen Linguist Lab to detect online hate speech: Evaluating online harms and digital rights violations

What is the @SentinelProject? How can we work together to harness #Hatebase as an effective tool to respond to threats posed by hate speech online?

@SentinelProject "We found that automation is a better step in terms of monitoring content. Yet we encountered challenges with linguistic coverage, since we monitor 90+ languages."

@RaashiSxn

@DrewBoyd: #Hatebase will generate a percentage output on whether certain content is considered hate speech. The scale ranges from 0 to 100%, so it's not a black and white system.

@SentinelProject @DrewBoyd :A lot of people conflate the concepts of freedom of speech and social media expression. For example, as a private company @facebook can impose certain limitations. It can have very specific terms of services and that would be legitimate. (👇)

@SentinelProject @DrewBoyd @Facebook @DrewBoyd: But this can create a very localized underground community to seek more extreme platforms that turns into an echo chamber, rather than a space for the free exchange of ideas.

"The idea is to have this system where the gathered hate speech is ultimately given one final review, or consistent oversight by human." DB

"#hatebase is in many ways research oriented more than a finished project that will go about filtering other people's content. We're working to discover answers around things that do and don't work." DB

"The finding of something not being possible is also a finding. We don't confidently say that this will filter hate speech online, but it's a way to better understand the tools, the systems, and the ecosystem, and to try to find our best way froward." DB

@ctuckwood on misinformation and hate speech:

They are still discussed as completely separate phenomena, but they're closely related, and there's overlap between the two of them.

i.e. If someone is expressing hostility towards a certain group of people they would share misinformation to justify their hate.
@ctuckwood

@ctuckwood: "Someone who's just reading negative things about a group doesn't necessarily pick up a weapon and attacks them. But when there's misinformation, it can actually have that triggering effect. We've found that misinformation could prompt people to take action."