DETECTION: To intervene fast against hate speech, automated classification is needed. But existing classifiers, e.g. Perspective API, are unreliable and do not catch the actual problem. They focus on slurs, not intolerance and are political biased (in all directions) [3/13]
-
Prikaži ovu nit
-
INTERVENTION: Even if detected, most tested interventions against hateful individuals have small effects, if any. But: The succes of counter-speech increases if it comes from leaders. If SoMe debates are to become better, our leaders need to orchestrate it [4/13]
1 reply 0 proslijeđenih tweetova 3 korisnika označavaju da im se sviđaPrikaži ovu nit -
BYSTANDERS: Because of the above, a focus on hateful people is not enough. Education should focus on how to be a good bystander. If you see something, interfere. If you fear retaliation, reach out to victims in private. Help them cope [5/13]
1 reply 0 proslijeđenih tweetova 4 korisnika označavaju da im se sviđaPrikaži ovu nit -
REPORTING: As part of this, better reporting systems needs to be developed and a key part of digital education should focus on how and when to report. Humans, not algorithms, is our first line of defense [6/13]
1 reply 0 proslijeđenih tweetova 3 korisnika označavaju da im se sviđaPrikaži ovu nit -
IDENTIFICATION: Current legislation already contains the needed elements to respond to hate speech and threats. The key problem: We cannot identify the posters. A solution is to require ID to get accounts [7/13]
1 reply 0 proslijeđenih tweetova 2 korisnika označavaju da im se sviđaPrikaži ovu nit -
REGULATION: Regulation is needed but should focus on process, not content: (1) content-regulation violates freedom of speech; (2) tech giants are the world's premier distributors of info and content-regulation gives them unbelievable power to shape the minds of everyone [8/13]
0 proslijeđenih tweetova 6 korisnika označava da im se sviđaPrikaži ovu nit -
ALGORITHMS: A key focus in process-regulation should be on the nature of algorithms. What should they prioritise and what should they downplay? Of key importance in the fight against misinformation. Still, to be effective against hate speech, we need those classifiers [9/13]
1 reply 0 proslijeđenih tweetova 5 korisnika označava da im se sviđaPrikaži ovu nit -
DATA ACCESS: To be able to have open discussions about process-regulation, it is of key importance that third parties have access to SoMe data and this should in itself to be a key focus for regulation [10/13]
1 reply 1 proslijeđeni tweet 9 korisnika označava da im se sviđaPrikaži ovu nit -
WIDER SOCIETY: It is not tech industry's responsibility that the world is a mess. Problems on the platforms reflect wider societal problems and our political leaders are responsible for solving those. But for them to do so, it is vital to civilize the online dialogue [11/13]
1 reply 0 proslijeđenih tweetova 7 korisnika označava da im se sviđaPrikaži ovu nit -
COLLABORATION: For this challenge, researchers, industry and policy-makers need each other. Hubs for information-sharing needs to be established. Researchers need to disseminate knowledge faster and practitioners need to help researchers identify the key problems [12/13]
1 reply 0 proslijeđenih tweetova 4 korisnika označavaju da im se sviđaPrikaži ovu nit
ROPH PROJECT: Civilizing online dialogue constitutes the core focus of @ROPHproject. Over the next years, we will test the impact of some of the above in the lab and in the wild. If you work with similar problems, let us know and we can make the world better together [13/13]
-
-
Odgovor korisnicima @M_B_Petersen @ROPHproject
This is excellent (both the project and the summary). Couple of questions. Will you address (1) different strategies and value orientations inherent to various parties? (e.g. max. eyeballs vs. accountability); (2) black-boxing inherent to stacked algorithms? (e.g. coral reefing)
1 reply 0 proslijeđenih tweetova 1 korisnik označava da mu se sviđa -
Odgovor korisnicima @GerritsLasse @ROPHproject
We should talk - would love to hear more about the thoughts underlying those questions!
0 replies 0 proslijeđenih tweetova 0 korisnika označava da im se sviđa
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.