Within the past year, Pearson has launched Aida, a mobile app that uses various AI forms to offer a deeply individualized educational experience. Marketed as the first educational app that doesn’t just point users to the answer, but also assists learners in discovering how to arrive there, Aida helps students tackle college-level calculus – typically one of the first courses college students must grapple with in order to pursue a career in STEM. Aida provides real-time feedback and targeted recommendations. As students increasingly interact with the app’s content, Aida’s tips and suggestions become more personalized to the user. Perhaps Aida’s most critical feature is its capacity to give targeted tips on users’ work. Aida can analyze photos of handwritten work or answers typed into the app. The app can give specific feedback on each line of work – informing students what they got right and offering resources to hone in on specific skills where students could use additional practice. Pearson is a publicly-traded company that earned $4,679.5 million in 2019.
Chatterbox enables language learning for professionals through AI-backed self-teaching materials and consistent practice with algorithmically-chosen live coaches. The platform pairs users with native language tutors – all of whom are refugees with professional backgrounds. Tutors and students are matched based on common professional histories and interests. For example, Congolese medical doctors are paired to teach French to medical aid workers, and Syrian engineers are paired to teach Arabic to engineers. While Chatterbox’s users are learning new languages, their teachers are able to earn a living, support their professional images and augment their employability. Chatterbox has raised $0.3 million to date.
Using proprietary natural language processing (NLP) technology, Cognii's Virtual Learning Assistant (VLA) is the sole electronic assistant (think Apple’s Siri, Amazon’s Alexa, etc.) that is geared toward educational dialogue. Educational dialogue, in this context, refers to the typical conversations a student would have with a teacher or tutor. The company's NLP technology allows the platform to assess the contextual semantics, syntax and conceptual hierarchy in students' written responses to open-answer assessments. In this way, Cognii can evaluate the correctness of responses, as well as identify the particular ideas in the response. Cognii then gives feedback to users much like a human teacher would. Cognii uses AI to support free-response assessments, smart tutoring, individualized deeper learning and robust analytics. In a regulated research study, Cognii was evaluated as 96% as accurate as humans in assessing concise essays. By employing machine learning and data mining, the accuracy of Cognii's evaluations increases over time. Cognii has raised $0.1 million to date.
A bonafide unicorn (i.e. a startup valued at over $1 billion), Duolingo is a language learning platform whose 300 million active users carry out billions of language-learning activities each month. Using machine learning and its vast data repository, Duolingo has engineered a statistical blueprint, known as "half-life regression" of how long users are able to recall the novel words they memorize before they begin forgetting and, therefore, require a reminder. Duolingo has developed the regression by assessing the errors of millions of users to forecast the "half-life" of each word in a human's long-term memory. Duolingo has raised $148.3 million to date.
Examity applies proprietary AI proctoring to online testing. The platform assesses test takers' identities and uses AI and machine learning to discover indications of cheating. The identity assessment is composed of three phases. First, the test-taker shows an official form of identification and submits a real-time image through the webcam. These images are evaluated against the ID on file. Next, test-takers respond to "challenge" questions. At last, the test-taker must provide a digital signature that assesses the "keystroke cadence" (i.e. the rhythm with which an individual types) of the test-taker and matches it against the test-taker's stored digital signature. Examity's automated proctoring technology is put into play when the identity assessment concludes. The proctoring technology oversees the exam from beginning to end through predictive analytics and video review. By "flagging" signs of cheating and suspicious behavior, Examity enables a faster human review process. Examity has raised $121 million to date.
By: Remi Nathanson