2026年10月15日（木）マイページ
← 記事にもどる 中学生版高校生版 Word

「印刷する」→ 送信先で「PDFに保存」を選ぶと PDF になります。解答は最後のページです。

原田英語の時事英語ジュニア｜英語ニュース ワークシート（高校生レベル）

Class ____ No. ____ Name ____________________

AIで勉強すると粘り強さが落ちる？10分の利用で自力の正答率が下がった実験

Leaning on AI for Answers May Weaken Persistence

音声・ゲームはこちらから

1Using an AI assistant for only about 10 minutes made people less accurate and quicker to give up once the tool was taken away, a new study has found. The research was presented this month at the Conference on Language Modeling, a major meeting on AI held in San Francisco. The team included researchers from Carnegie Mellon University, MIT, the University of Oxford and the University of California.

2The researchers ran three experiments with a total of 1,222 people in the United States who signed up online. In each one, participants were randomly divided into two groups. One group could use a chatbot during a practice stage, and the other could not. Then both groups took a short test with no AI at all.

3In the first experiment, which used fraction problems, the AI group solved 57 percent of the test questions, compared with 73 percent for the group that never had AI. The AI group also skipped 20 percent of the questions, almost twice the rate of the other group. A third experiment with reading questions from SAT practice tests showed the same pattern, as the AI group skipped 8 percent of the questions, compared with 1 percent for the others.

4How people used the AI seemed to matter. In the second experiment, 61 percent of the AI users said they had mainly asked it for direct answers, and this group scored lower than the group without AI. Those who mostly asked for hints or explanations did about as well as people who never had AI. The researchers noted that this part of the study shows a link, not proof of cause.

5The researchers said AI tools should act more like tutors that guide people than machines that hand out answers. The study measured only short tasks, so it did not show how long the effect lasts.

Words　単語

accurateB1形正確なcorrect and with no mistakes
presentedA1動発表したshowed new work to a group of people
experimentsB1名実験tests done to find out something new
participantsB1名参加者people who take part in something
randomly外副無作為に、ランダムにby chance, without a plan or order
compared withA2熟～と比べてwhen you look at it next to something else
patternB1名傾向、型a way something happens again and again
directA2形直接の、そのものずばりのgoing straight to the point
proofB1名証拠、証明facts that show something is true
tutorsB2名家庭教師、個人指導の先生teachers who help one student at a time

単語のレベル：CEFR-J（『CEFR-J Wordlist Version 1.6』 東京外国語大学投野由紀夫研究室. （URL: https://www.cefr-j.org/download.html より 2026年8月ダウンロード））

1. Reading Quiz　本文の内容に合うものを選ぼう

  1. Where was the research presented?

    A. At a school in JapanB. At a conference on AI in San FranciscoC. On a TV showD. At a math contest

  2. How were participants put into groups?

    A. By their ageB. By their test scoresC. They chose their own group.D. Randomly

  3. In the first experiment, how many questions did the AI group skip?

    A. 20 percentB. 8 percentC. 57 percentD. 1 percent

  4. Which AI users scored lower than the group without AI?

    A. Those who asked for hintsB. Those who never opened the AIC. Those who mainly asked for direct answersD. Those who asked for explanations

  5. What does the study NOT show?

    A. How long the effect lastsB. That the AI group skipped moreC. That 1,222 people took partD. That hint users did about as well

2. True or False　正しければ T、違っていれば F

  1. (     ) The study was done with high school students in Japan.

  2. (     ) Both groups took the final test without AI.

  3. (     ) In the SAT reading experiment, the AI group skipped more questions than the other group.

  4. (     ) The researchers said they proved that asking for answers causes lower scores.

3. Fill in the Blanks　下の語から選んで空らんをうめよう（形を変えずに）

participantspresentedpatternaccuratedirectexperiments

  1. Using an AI assistant for only about 10 minutes made people less _____ and quicker to give up once the tool was taken away, a new study has found.

  2. In each one, _____ were randomly divided into two groups.

  3. In the second experiment, 61 percent of the AI users said they had mainly asked it for _____ answers, and this group scored lower than the group without AI.

  4. The researchers ran three _____ with a total of 1,222 people in the United States who signed up online.

  5. A third experiment with reading questions from SAT practice tests showed the same _____, as the AI group skipped 8 percent of the questions, compared with 1 percent for the others.

  6. The research was _____ this month at the Conference on Language Modeling, a major meeting on AI held in San Francisco.

4. Word Order　語を並べかえて正しい文にしよう

  1. ( the / matter / people / used / How / AI / seemed / to ) .

  2. ( all / Then / test / groups / at / AI / both / no / with / took / a / short ) .

  3. ( randomly / each / divided / were / groups / participants / In / two / one, / into ) .

5. Think and Write　英語で答えてみよう

  1. Should schools allow students to use AI for homework? Give reasons for your opinion.（学校は、生徒が宿題にAIを使うことを認めるべきだと思いますか。理由をあげて意見を言おう。）

  2. What rules would you make for yourself when you use AI to study?（AIを使って勉強するとき、自分にどんなルールを作りますか。）

Answers　解答

1. (1) B　(2) D　(3) A　(4) C　(5) A

2. (1) F　(2) T　(3) T　(4) F

3. (1) accurate　(2) participants　(3) direct　(4) experiments　(5) pattern　(6) presented

4.

  1. How people used the AI seemed to matter.
  2. Then both groups took a short test with no AI at all.
  3. In each one, participants were randomly divided into two groups.

全文訳

1AIアシスタントをわずか10分ほど使っただけで、そのAIを取り上げられたあと、正確さが下がり、早くあきらめるようになることが、新しい研究でわかりました。この研究は今月、サンフランシスコで開かれたAIの大きな学会「言語モデル学会（COLM）」で発表されました。研究チームには、カーネギーメロン大学、マサチューセッツ工科大学（MIT）、オックスフォード大学、カリフォルニア大学の研究者が入っていました。

2研究者たちは、インターネットで参加したアメリカの合計1222人で、3つの実験を行いました。どの実験でも、参加者はくじ引きのようにランダムに2つのグループに分けられました。1つのグループは練習の段階でチャットボットを使えましたが、もう1つのグループは使えませんでした。そのあと、両方のグループがAIなしで短いテストを受けました。

3分数の問題を使った最初の実験では、テストの正答率がAIを使ったグループで57％、一度もAIを使わなかったグループで73％でした。AIを使ったグループは問題の20％を飛ばし、これはもう1つのグループのほぼ2倍でした。アメリカの大学入試SATの練習問題から読解問題を使った3つ目の実験でも同じ傾向が見られ、AIを使ったグループは問題の8％を飛ばしました。もう1つのグループは1％でした。

4AIをどう使ったかが、結果に関係しているようでした。2つ目の実験では、AIを使った人の61％が主に答えそのものを聞いたと答え、このグループはAIなしのグループより点数が低くなりました。主にヒントや説明を聞いた人は、一度もAIを使わなかった人とほぼ同じくらいできました。研究者たちは、この部分は「関係」を示すもので、原因だと証明したものではないと述べています。

5研究者たちは、AIは答えを渡す機械というより、人を導く家庭教師のようにふるまうべきだと話しています。この研究は短い課題しか調べていないため、この影響がどのくらい続くのかはわかっていません。

文法ポイント：使役動詞 make + 人 + 形容詞（人を～にする）

make + 目的語 + 形容詞 で「（人・もの）を～の状態にする」。本文の Using an AI assistant for only about 10 minutes made people less accurate and quicker to give up. は「AIを10分ほど使うことが、人を（前より）不正確に、あきらめやすくした」。主語が「～すること（動名詞）」になっている点にも注意。

出典：arXiv（Liu, Christian ほか）、Tech Xplore、Science News Explores ほか（記事は原田英語の時事英語ジュニアのオリジナル）｜https://www.haradaeigo.com/jijieigo-junior/news/2026-10-15-ai-help-less-persistence/