Re: Man vs. Computer: Who Wins the Essay-Scoring Challenge?
From Education Week: Curriculum Matters 4-13-2012
“The results demonstrated that overall, automated essay scoring was capable of producing scores similar to human scores for extended-response writing items with equal performance for both source-based and traditional writing genre,” says the study.
This is fascinating. Artificial intelligence has already progressed to the level that it’s capable of competing with the minds of college professors at assessing (if only through limited scope written essays for now…) the learning and performance of students. Within a few years of being established in colleges it will grow the capability to assess student learning through verbal conversation (like WATSON) and visual observation, and begin working its way down to preschools.
IBM’S WATSON – Natural Language
AI that won TV’s Jeopardy championship
How many years before such AI systems expand to include birth hospitals, doctors, parent profiles, and preschools? Imagine how different education will become once AI assessment is developing life-long learning profiles on all of us – when AI is continually learning about each us from birth through every grade in school (and promotion or advancement at work). This is not science fiction, it’s the inevitable application of AI everywhere that it’s profitable to apply it. It might be 10 or 15 years before it trickles down through the all the levels of school but, given the enormous costs involved in education, it’s a no-brainer that it will.
(note: our page “The Challenge of Change” provides some important context to what follows)
Some considerations…
Will we need fewer professors when professors are more efficient? How many professors might this replace?
Currently educators learn a lot about their students by reading their essays. How will AI assessments change how educators learn about their students?
How do we ensure the data and personal profiles developed by such a system aren’t used unethically?
Which ideological views, biases, and assumptions about human nature and development will inform how the AI interprets student assessments?
IN 10 YEARS – MILLIONS OF TIMES MORE POWERFUL THAN TODAY
PALS: Personal Artificial Learning Stewards – as intelligent devices (“the descendants of PCs, Laptops, Netbooks, Tablets, and smartphones) become cheaper, more powerful, and even more ubiquitous, it will become possible for each of us to have a personalized artificially intelligent learning support systems that is capable of responding to any question we can articulate and that is available 24×7 through any one of our numerous devices.
A THOUGHT EXPERIMENT ABOUT PALS
With Watson like intelligence participating in and assessing our every interaction, these devices will eventually develop to the point that they become more than just responsive learning aids, they will become proactive teachers. I hope that by then we will have revisioned the mission of education. If these machines have an agenda or ethic other than ‘stewarding the health of our learning’ we are going to be in big trouble.
Further reading:
From Here to Implicity – The Public Learning System
PLEASE… We can’t sustain our efforts without your help. If you think we are on the right track, or even one that should be given consideration, please help us. Please forward our posts as widely as you can. Please contribute whatever you can, $5, $10… whatever you can (see “Donation” button in sidebar). Arrange for us to work with your school or organization. Your help can make a big difference in whether our work succeeds. Thank you.
No comments yet.