Disability-Centered AI And Ethics MOOC

AI algorithms can be used to augment smart wheelchairs, walking sticks, geolocation and city tools, bionic and rehabilitation technologies, turn sign language into text, or visual objects into transcriptions, support cognitive impairments - through expressive computing and social robotics. These algorithms also fuel a range of systems used by the general population (“assistive pretext”):

Despite the significant possibilities of AI, disability is not a monolith, but a social spectrum that requires special attention and may pose challenges for algorithms associated with proper recognition, analysis, predictions and outcomes.

In particular, Al systems can make errors or lack accessibility towards individuals with sensory, cognitive or physical impairments in the context of the workplaces, educational, public or law-enforcement scenarios, including:

Individuals with facial differences or asymmetry, craniofacial syndromes

Different gestures, gesticulation, posture;

Speech impairment, specific communication styles;

Used assistive devices or technologies;

Negative connotations for semantics associated with disability keywords and phrases (e.g. language-based systems);

Be less accurate towards “invisible“ disabilities, women (“gender-blind“) or particular populations.

MOOC explores:

how different AI systems and ontologies, including expressive computing, may interact with different spectrums;

investigate historical and statistical contexts of errors and distortions; 

explore approaches to data, models, systems oversight;

explore policies and frameworks to systems categorization, scenarios, risks and oversight.