Sunday, April 26, 2020

Are trolls born or can we all be one?

Social media are the hunting grounds of trolls and few news media comments sections escape their attention, leading in many cases to the withdrawal of such sections. There are of course different degrees of trolldom from the infamous troll factories who make money deliberately sabotaging political discussion to individuals with anger management issues. I often wonder how people can thrive on such negativity and enjoy insulting, threatening and winding up other people, but it seems they do. But is there a little troll inside us all, waiting for the right cues to rise to the surface?

There is of course research on troll behaviour and I can recommend this article from researchers at Stanford and Cornell Universities, Anyone Can Become a Troll: Causes of Trolling Behavior in Online Discussions. They carried out an extensive survey of comments on CNN news stories as well as conducting a simulated online discussion to see what patterns of troll behaviour emerged and whether such behaviour is infectious. It seems that trolls are not just born, we all risk getting sucked into a certain element of trolling if the right conditions are met.
While prior work suggests that trolling behavior is confined to a vocal and antisocial minority, we demonstrate that ordinary people can engage in such behavior as well. We propose two primary trigger mechanisms: the individual’s mood, and the surrounding context of a discussion (e.g., exposure to prior trolling behavior). Through an experiment simulating an online discussion, we find that both negative mood and seeing troll posts by others significantly increases the probability of a user trolling, and together double this probability.
I think many of us will admit to firing off the occasional sarcastic retort on Twitter or Facebook especially when someone has made an unfair or abusive comment. We might even get involved in a mud slinging session that we later regret. However for me a troll is someone who joins a discussion with the deliberate intent to sabotage a discussion or to bully particular members of the group. For me, trolling is bullying, taking pleasure in hurting others. In many online communities there are volunteers who devote their evenings and even nights to patrolling the digital streets and trying to defuse potentially dangerous situations. By immediately intervening in a thread that is going out of control they can defuse the situation but only when they are dealing with people who are simply overreacting or in a foul mood. The genuine trolls are almost impossible to stop since they thrive on confrontation and the only way to deal with them is by ejecting them from the community.

In such well-managed discussion groups with clear guidelines, timely moderation and real name policy, trolling can be kept at a minimum and this leads to a supportive and tolerant community. It requires great sensitivity on the part of the administrators to be able to distinguish between people who are just having a bad day and go too far once in a while and genuinely malevolent trolls. In many groups an administrator will try to reason with the offender, preferably in a private chat, and this often results in that person deleting their offensive comments and agreeing to abide by the group rules in the future. Many have a policy of three strikes and you're out, giving warnings and then if the abuse continues the offender is ejected. In some cases instant ejection is the only way to deal with abuse but for many a certain amount of leeway might help. The article ends by warning against simply banning everyone who steps out of line and recommends greater flexibility.
Trolling stems from both innate and situational factors – where prior work has discussed the former, this work focuses on the latter, and reveals that both mood and discussion context affect trolling behavior. This suggests the importance of different design affordances to manage either type of trolling. Rather than banning all users who troll and violate community norms, also considering measures that mitigate the situational factors that lead to trolling may better reflect the reality of how trolling occurs.
Cheng, J., Bernstein, M., Danescu-Niculescu-Mizil, C., Leskovec, J. (2017) Anyone Can Become a Troll: Causes of Trolling Behavior in Online Discussions. CSCW '17: Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing, February 2017. Pages 1217–1230

Wednesday, April 15, 2020

Transition to online teaching - so far so good but who are we leaving behind?

Photo by Tim Gouw from Pexels
With the vast majority of schools and universities in the world teaching online there is plenty scope for research into how well or not the transition has been managed and what conclusions can be drawn from the experience. Already there are survey results indicating that most students are satisfied with the online solutions being offered, given the exceptional circumstances prevailing. However that should not lead to complacency for behind the positive indications there are also students who are being excluded.

One new survey, of law students at the University of Oslo, is described in an article in Mirage News, Digital teaching provides good learning outcomes, lightning-fast research shows. A week after moving to online teaching 175 students were asked about their experience of the transition and although this is an extremely limited survey it does reveal some interesting tendencies (the published article is available in Norwegian, Første uken med digital undervisning i koronatiden: Jusstudenters erfaring). 60% of the students stated that the online teaching was as good as or even better than the campus teaching and the teachers express surprise at how well the transition was received. One teacher explains:
My main thought is that the findings indicate that the students across the faculty have positive experiences from the first week of digital teaching. Although there is of course some criticism here and there, I had really expected far greater dissatisfaction, especially considering that this evaluation applies only to the first week. At the same time, we need to be a little cautious about the use of these findings, as they are only based on the first week of digital teaching.
At the same time there was a significant number of students who were less satisfied and this was mostly due to other technical and social factors.
A substantial minority of students are struggling with childcare, illness, poor internet connection, lack of workspace at home and lack of contact with teachers and fellow students. There is a strong correlation between the number of challenges and how the student evaluates their learning experience and study situation.
Most of these factors are beyond the control of the university but are critical to students' academic success, not just in exceptional times such as these. One uncomfortable truth is that university education is still largely geared to the traditional student model who are able to live and study on campus, do not need to combine study with work and do not have families to care for. University study whether on campus or online is designed for the traditional notion of a student, as described in an article on The ConversationUniversity study is designed for the privileged – students from disadvantaged backgrounds suffer. This becomes even more true in the exceptional circumstances today. Those who cope best with online learning are those who are already have good study skills, have access to fast internet connections, own the right devices and have a social environment that allows them to study in peace. For others simply finding a place to study in peace is a challenge, especially when even places like public libraries are closed. 
Additionally, flexibility may make it harder to carve out time for studies. Scheduled on-campus blocs of time allow students to focus, which they may not be able to do when faced with the immediacy of children or younger siblings not attending school. Many students are working in retail, meaning they may be their family’s only source of income as parents are unable to work. The risk is that the pandemic exacerbates existing inequities and makes it even harder for these students to engage with their studies.
These themes are further developed in a blog post by Tharindu Liyanagunawardena (UCEM, London) Online learning in challenging times. In the rush to switch to online mode it is all too easy to forget accessibility issues and not take the students' home situation into consideration. We cannot, of course, remedy their social situation but should at least try to ensure that our platforms and tools are as accessible as possible. For example we can ensure there are transcripts of the video lectures for those without access to broadband or who prefer to read than to listen. Maybe an asynchronous activity is easier to access and participate in than a synchronous video meeting? Tharindu gives further examples:
For example, if a student with a hearing disability was supported by a note-taker in class how could we support this student now that we have moved to online lectures? Or now that most overseas students have gone home to their countries, can we conduct online classes at the same time and expect them to be present despite the time differences? What if the technology we adopt is barred in some countries where our students reside?
It is only natural that in the scramble to move online some issues are forgotten or omitted due to lack of time and resources but as we move on these issues need to be addressed. In all situations we need to take time to consider who we are excluding or limiting and try to find the right balance.

Sunday, April 5, 2020

Zoom - a question of trust

Photo by ThisIsEngineering from Pexels
Since my last post there has been increasing media attention on Zoom, not just on the phenomenon of zoombombing that I experienced but also on various serious security and privacy issues. The platform has become almost default for millions of teachers in schools and universities around the world, increasing its usage from about 10 million users in December to over 200 million in March. But such extreme usage has revealed that privacy and security were not top of the company's priorities according to an article in NPRA Must For Millions, Zoom Has A Dark Side — And An FBI Warning.
"Things you just would like to have in a chat and video application — strong encryption, strong privacy controls, strong security — just seem to be completely missing," said Patrick Wardle, a security researcher who previously worked at the National Security Agency.
The article also tells of several unfortunate examples of zoombombing including a doctoral thesis defence that got hijacked and a meeting of a branch of Alcoholics Anonymous, prompting even a warning from the FBI. Zoom has been working hard at calming fears and they are now prioritising security issues while putting new feature development on hold (see the Zoom message to users from 1 April). It was also revealed that the company have been sharing user data with Facebook and LinkedIn (see Mashable), something that they admitted and claimed was a mistake. Furthermore, a bug was found that enabled hackers to access users' accounts (see Mashable). I can imagine that many of Zoom's staff have had little sleep during the past week.

I have been trying to adjust my Zoom settings and giving participants much less control than before. Many articles recommend using passwords for all meetings but I haven't gone that far yet. Since I'm often involved in webinars and open sessions in Zoom we usually want to reach a wide audience. I love the idea of people from different places, professions and areas of expertise getting together to discuss and exchange ideas and up till now that has been possible using different e-meeting platforms. Adding passwords and so on adds barriers to spontaneous participation and it is sad to lose that opportunity because of the destructive behaviour of a minority of idiots.

I'm sure Zoom will address all these issues and are promising regular updates on progress, but the central issue here is one of trust. The education sector works with children and young people whose privacy and integrity we have a duty to safeguard. We are also dependent on commercial platforms and tools that we assume also respect this duty and with whom there are special agreements adapted for the education sector. But if we find that there are loopholes in these agreements that trust is broken and we have to face the question of who we should trust in the future. I can almost understand that if you use a service that is labelled as free there will be a price in terms of how my data is used, but if you are paying a lot of money for a tailored educational solution then there should be very strict controls on encryption, data protection and so on. If these companies want to be in the education sector they have to be able to guarantee security and integrity. The alternative is for the education sector to run its own platforms in its own infrastructure and be in control of its own security. Not a very likely scenario given the costs but in today's world who knows what lies ahead.

Update: A good and balanced overview of the situation is an article, Zoom isn't malware, by three security experts.