| Educators know that teaching students
to write well is not easy. One problem is the amount of time needed to
{{U}}(1) {{/U}} large {{U}}(2) {{/U}} of work. So some companies
have developed computer programs. These can {{U}}(3) {{/U}} student
writing much more quickly than a human can. Writing tests can also cost
{{U}}(4) {{/U}} to {{U}}(5) {{/U}} by computer than by
paper-and-pencil. These computer systems are known as e-raters. They use artificial intelligence to think in a way {{U}}(6) {{/U}} teachers. In the state of Indiana, computer grading of a statewide {{U}}(7) {{/U}} test began with a {{U}}(8) {{/U}} of the system itself. For two years, both a computer and humans graded the student writing. Officials say there was almost no difference between the computer grades and {{U}}(9) {{/U}} given by the {{U}}(10) {{/U}} readers. Systems are also being used to grade writing in college classes. The computers {{U}}(11) {{/U}} a few hundred {{U}}(12) {{/U}} of student writing already graded by humans. Then the systems {{U}}(13) {{/U}} new essays {{U}}(14) {{/U}} those already examined. How do teachers feel about all this? Many say machines can {{U}}(15) {{/U}} the job {{U}}(16) {{/U}} people can. A computer can find spelling and grammar mistakes. {{U}}(17) {{/U}} these teachers say it can never really {{U}}(18) {{/U}} what a writer is trying to say. {{U}}(19) {{/U}} say a program cannot follow a thought or judge a humor {{U}}(20) {{/U}} understand a beautifully expressed idea. |