KNC13 OPENGOV | OPEN BS DETECTOR: Baffle-speak sensing and contextualization | Knight News Challenge

OPEN BS DETECTOR | Baffle-speak sensing and contextualization

Short URL:

Twitter: @OpenBSDetector

Public officials too often rely on rote phrases devoid of real meaning (baffle-speak) when answering questions of public interest. Open BS Detector spots the sound bites, logs, tracks and alerts about them, and offers useful information and context.

Citizens and journalists need to be able to more quickly and easily discern fact from fiction or misdirection in statements by their politicians and public officials, put those statements in context, and get accurate facts to make informed decisions and take action.

Open Baffle-Speak Detector will parse those comments and answers to questions, compare them to a historical record by the individual on a topic, and return statistical information, as well as verified facts to the individual seeking a better understanding of a subject and an official’s approach to it.

Open Baffle-Speak Detector will have three modes to enable people to identify and contextualize the accuracy, veracity and utility of officials’ statements:

  1. A browser plug-in for use on news articles, online transcripts and other text or Web pages.
  2. Voice-recognition/transcription and a Shazam-like mode for real-time assessment.
  3. A photo/facial recognition and/or augmented reality function that enables the user to easily identify the individual, their track record and overall baffle-speak score.

Our preliminary research discussing this idea among people in the open government movement, journalism, and technologists has been met with enthusiasm. The technology to do this exists or is in development in other arenas, and simply needs to be brought together for an end-to-end, easy experience rather than a patchwork that results in a laborious series of tasks, or a lack of capability in this particular sphere. We plan to collaborate with as many developers of existing projects, tools and technologies as possible.

The Truth Goggles, Lazy Truth, Super PAC App, and Churnalism projects are a few examples of similar ideas with different applications and approaches. They focus on parsing text or standardized data. None handle live, real-time input. Open Baffle-Speak Detector will stitch these approaches together for a robust, on-demand tool that works in live situations.

We have not commenced development on Open Baffle-Speak Detector, but have had initial contacts with people working on other and related projects. We will collaborate and build on Truth Goggles, Hyperaudio and other projects as much as possible to avoid duplicating work.

Saleem Khan: Project leader, journalist [editor and reporter, ex- CBC, Metro International, Toronto Star newspapers; chairman/director, Canadian Association of Journalists]; advisor, University of Toronto ThingTank Lab [Faculty of Information]; founder,

Collaborators & Advisors:
M. Boas: OpenNews Fellow 2012. Hyperaudio leader, jPlayer HTML5 media library project coordinator, open Web developer.
L. Gridinoc: OpenNews Fellow 2012. Creative technologist specializing in computational linguistics, semantic Web, and visual analytics.
P. Hunter: Over 20 years designing productive interactions between people and technology; expertise in speech recognition, software tools, and education; Fellow in the Leading by Design program at California College of the Arts; veteran of three start-up businesses; currently at Microsoft.
K. Kaushansky: Two decades specializing in speech recognition, voice user interface design, interactive audio experiences, speaker verification, and voice biometrics at startups and global technology firms, including Nortel and Microsoft. Currently at Jawbone.
K. Khan: User experience strategist and designer consulting to governments and Global 1000 corporations, OCAD University sLab advisor; leader of UXI, Canada’s largest UX professionals group.
H. Leson: Director of community engagement, Ushahidi; open source community developer, library and information technician.
M. Saniga, CA: Co-founder, near-realtime business intelligence/data insight generation software firm Quant Inc.; former finance director and manager at Cara, Dell.

We anticipate that Open Baffle-Speak Detector will gain interest and uptake among civic development and open government foundations, news organizations, and real-time intelligence companies and investors which would continue to fund development and custom or applications-specific versions.

Open Baffle Speak Detector is a tool that empowers citizens to identify when public officials give rote vs. real answers in their statements, and adds verified factual context.

Toronto, Ontario, Canada