1. Small ideas for ICRC
Sanjana Hattotuwa
ICT4Peace Foundation,TED Fellow
2. what’s new
• Ubiquity of two way communications
• Addressable peoples, even those who IDPs or refugees
• Both news generation and dissemination leverages new media
• Disintermediated models vs. traditional media model
• Citizens as producers
• Low resolution content broadcast on high definition media
3. what’s new
• Sous-veillance (observing from underneath, anchored to human security) in place
of, or in addition to, surveillance (often from centralised loci, anchored to national
security)
• Sous-veillance is crowd based intelligence, generally open data (though analysis
can be bounded). Surveillance ranges from sig int and psy ops to information
espionage, almost always bounded.
• Important to understand Arab Spring, and situational awareness in sudden onset
disasters
4. Focus on process, not just spikes
Narrow band over time adds richness, full spectrum adds context
Local language(s)
Culture
Local actors
Diaspora
Hagiography and myth
Identity and power
Partisan politics
Regional power blocs
Inequity
Demographics (Youth)
Civic media
Verbal storytelling
17. challenges
• Concept of failing forward missing. Everyone parading what worked, but
more imp to know - what failed, why?
• Heard first cursory mention of ethics amidst overwhelmingly technocratic
perspectives. Good. Need to flesh out.
• No recognition of (geo) politics and US strategic interests in use & availability
of tech. Compare Haiti, Pakistan & Myanmar in '08
• A bigger disaster than Haiti, Pakistan had comparably little of this tech,
volunteerism and focus. Why?
18. challenges
• Surprisingly everyone seems to believe crowdsourcing is good, and is only
used for good. Context, content, creator, consumer absent
• At risk of sounding Rumsfeldian, why don't we know what we should know?
Core datasets vital for community resilience and response
• Trust is mutable, relative, contextual, locally defined, gendered, framed by
identity, inter alia.
• Violence as a result of knowledge creation.
19. challenges
• Impartial, accurate coverage still vital, increasingly hard to ascertain
• Torrent of information. Trickle of knowledge.
• Veracity hard to determine
• Pace of technology development hard to keep pace with
20. • Nature of violence, partisan bias, citizenship, governance structures, public
institutions heavily influence crowdsourcing.
• Crowdsourced HR or election violations mapping with volunteers from
perpetrator party/tribe/ethnicity? Proceed with caution
• Volunteerism undergirding stand-by crowdsourcing good, but what about
CPE's, where personal bias can deeply influence curation?
• Related to last tweet, volunteerism works better for sudden onset natural
disasters, which are also mediagenic
enduring challenges with crisismapping and
crowdsourcing
21. how and who do we trust?
abduction of a gay girl of damascus. or so we thought.
Tom MacMaster, 40 year old American
http://damascusgaygirl.blogspot.com Jelena Lecic, of London
22. A lesbian in Damascus
And other tall tales
Disinformation
Misinformation
Partial accounts
Gaming the system
Gender imbalance (e.g. rape reports in DRC)
Lack of access leads to challenges in verification
Multiple retweets mistaken for authenticity
Anonymity online (esp. post-Norwegian terrorist attack)
Machine translation / Lack of translation
Little or no direct access
Trauma
Anxiety
Fear
Persecution
Network infiltration and disruption
Trust perceptions and authority markers
Bias in mainstream media
Bias in citizen media
23. filter bubbles
• "A Squirrel Dying InYour FrontYard May Be More RelevantToYour Interests Right NowThan
People Dying In Africa", Mark Zuckerberg, creator of Facebook
• Human gatekeepers being replaced by algorithmic gatekeepers.
• A new, pervasive, almost invisible, systemic filtering?
http://www.ted.com/talks/eli_pariser_beware_online_filter_bubbles.html
24. filtering to counter filter bubbles
• Ushahidi SwiftRiver | http://ushahidi.com/products/swiftriver-platform
http://www.youtube.com/watch?v=Tb0Gs7vtrgk
SwiftRiver is a platform that helps people make sense
of a lot of information in a short amount of time.
In practice, SwiftRiver enables the filtering and
verification of real-time data from channels like
Twitter, SMS, Email and RSS feeds.
25. two key effects of information overload
• Continuous partial attention, Linda Stone, Microsoft,
1997. With continuous partial attention we keep the
top level item in focus and scan the periphery in case
something more important emerges.
• The immediate altruistic response rapidly diminishes
over time (Melissa Brown, associate director of
research at the Center on Philanthropy at Indiana
University, 2010) Our brains release congratulatory hits
of dopamine when we engage in selfless behaviour —
which we’re moved to do the instant we witness
something awful.
28. CiM drivers from other domains
• Music industry (pattern based search, e.g. Pandora’s technical + human indexing), social
networking (group collaboration,e.g. LinkedIn, Facebook), social networking search (e.g.
Grepling), mobile phone apps (e.g. Guardly), marketing engines (e.g. adaptive persuasion
profiling), digital forensics (e.g. hyperspectral imaging with UAVs), ground truth profiling (e.g.
UNOSAT images on Sri Lanka) many sourcing for situational awareness (e.g. Microsoft
Photosynth), Open Data Initiatives (e.g. British, US govt’s,World Bank), visualisation (e.g.
Infomous)
29. take home
• Think beyond text. Online is not print.
• Think beyond prose. Online can be satire, verse, haiku!
• Think of photos, audio, video. Rich media tells stories, adds context.
• Think of SMS and crowd-sourcing, the audience are the producers.
• Don’t suggest you know everything. Use the community to add value to story.
• Link to other stories online, they add value.