EigoPro 1級 読解 — The Algorithmic Mirror: When Crime Forecasting Reproduces Its Own Premises

Predictive policing rests on a seductive proposition: that the dispassionate aggregation of historical crime data can pinpoint where offences will next occur, allowing finite resources to be deployed with surgical efficiency. Yet this promise founders on a foundational fallacy. The data such systems ingest are not a neutral record of where crime happens, but a record of where policing has historically concentrated. Arrests in heavily patrolled neighbourhoods generate dense data; comparable offences in under-policed districts go unrecorded. The algorithm, blind to this asymmetry, mistakes enforcement patterns for criminality itself. The consequences are insidious precisely because they masquerade as objectivity. When a model flags a district as high-risk, officers are dispatched there, where their very presence yields more arrests, which are duly fed back into the system as fresh confirmation. This recursive validation manufactures a self-fulfilling prophecy: the prediction engineers the reality it purports merely to anticipate. Historical bias, far from being expunged by quantification, is laundered through it, emerging with the unimpeachable veneer of mathematical neutrality. Defenders counter that algorithms are inherently impartial, free of the prejudices that taint human judgement. But this conflates the absence of malice with the absence of bias. A model trained on skewed inputs will faithfully reproduce, and indeed amplify, the very distortions it inherits. The machine does not deliberate; it extrapolates. To invoke its supposed neutrality is to mistake the mechanism for an arbiter, and to forget that every threshold, weighting and category embedded in the code reflects a human decision about what counts as risk and whose conduct merits suspicion. What is most pernicious is the rhetorical armour the term 'data-driven' confers. It transmutes a contestable policy choice into an apparently empirical inevitability, insulating it from democratic scrutiny and disarming critics who would otherwise demand justification. Few elected officials care to be seen disputing the cold verdict of a computer. The remedy is not to abandon analysis but to interrogate the provenance of the data and the assumptions encoded within. Until then, such systems will continue to entrench the inequities they were heralded to dissolve, dressing prejudice in the borrowed authority of science.

5問合格 60%以上普通

新規登録 または ログイン が必要です

この単元の問題を3問、解説つきで公開しています

1級 · 読解 · 全5

1

Which statement best captures the central argument of the passage?

  • Predictive policing reproduces historical bias by treating records of past enforcement as if they were objective measures of crime.
  • Predictive policing fails chiefly because the algorithms it relies on are too rudimentary to process large volumes of crime data accurately.
  • Predictive policing should be embraced because its mathematical impartiality eliminates the prejudices inherent in human officers' decisions.
  • Predictive policing is ineffective only in neighbourhoods that have historically been under-policed and lack sufficient recorded data.
Phrase

mistakes enforcement patterns for criminality itself

1級読解主旨把握論旨
ひとことで

本文の核心は、過去の取り締まりの記録を客観的な犯罪指標と取り違えることでバイアスが再生産される点にある。

選択肢の解説Choice Analysis

Predictive policing reproduces historical bias by treating records of past enforcement as if they were objective measures of crime.

正解

筆者は、システムが取り込むデータは犯罪の発生場所ではなく警察が歴史的に集中してきた場所の記録だと述べる。アルゴリズムはこの非対称性に盲目で「mistakes enforcement patterns for criminality itself」、すなわち取り締まりのパターンを犯罪そのものと取り違える。これにより歴史的偏見が再生産されるというのが論旨の中心である。

  • Predictive policing fails chiefly because the algorithms it relies on are too rudimentary to process large volumes of crime data accurately.

    = アルゴリズムが原始的すぎて大量データを処理できないという主張 ━ 本文の批判はデータ処理能力の不足ではなく、データそのものの偏りにある。処理性能の問題には触れていない。。

  • Predictive policing should be embraced because its mathematical impartiality eliminates the prejudices inherent in human officers' decisions.

    = 数学的公平性が人間の偏見を排除するため推進すべきという主張 ━ 本文はこの擁護論を明確に反論しており、『conflates the absence of malice with the absence of bias』と退けている。。

  • Predictive policing is ineffective only in neighbourhoods that have historically been under-policed and lack sufficient recorded data.

    = 取り締まりが手薄な地域でのみ無効という主張 ━ 本文はむしろ過剰に取り締まられた地域でデータが密集し悪循環が生じると述べており、無効の範囲を特定地域に限定していない。。

Collocation
パターン意味
rest on a propositionある前提に立脚するThe theory rests on a proposition that markets self-correct.
mistake X for YXをYと取り違えるCritics mistake confidence for competence.
feed back into〜に再び供給されるThe output is fed back into the original dataset.
self-fulfilling prophecy自己成就的予言Pessimism can become a self-fulfilling prophecy.
insulate from scrutiny精査から遮断するThe clause insulates the deal from public scrutiny.
例文
  • The forecast mistakes correlation for causation.

    その予測は相関を因果と取り違えている。

  • Dense data accumulate only where attention is concentrated.

    注目が集中する場所にのみデータが密に蓄積する。

  • The model reproduces the very bias it inherited.

    そのモデルは受け継いだまさにその偏見を再生産する。

  • Objectivity is claimed but never demonstrated.

    客観性が主張されるが決して証明されない。

💡Tip · 覚えるコツ

GIST問題では、選択肢が本文の一部分だけを誇張したり、筆者が反論している立場を正解に見せかける罠が多い。筆者の『評価のスタンス』全体と一致するものを選ぶ。

2

According to the passage, why do heavily patrolled neighbourhoods generate dense crime data?

  • A greater police presence there leads to more arrests, which are recorded as data.
  • The algorithms are specifically calibrated to collect more information from those areas.
  • Residents of those neighbourhoods are more inclined to report offences to the authorities.
  • Such neighbourhoods inherently experience a higher genuine rate of criminal activity.
Phrase

Arrests in heavily patrolled neighbourhoods generate dense data

1級読解詳細把握
ひとことで

密なデータは警察の集中配備による逮捕の増加から生じる、と本文は明示する。

選択肢の解説Choice Analysis

A greater police presence there leads to more arrests, which are recorded as data.

正解

本文は『Arrests in heavily patrolled neighbourhoods generate dense data』と述べ、対照的に取り締まりの手薄な地区では同様の犯罪が記録されないとする。つまりデータの密度は実際の犯罪率ではなく警察の配備量を反映する。これが非対称性の核心である。

  • Residents of those neighbourhoods are more inclined to report offences to the authorities.

    = 住民が当局に犯罪を報告しやすいから ━ 本文は住民の通報傾向に一切言及しておらず、密度の原因を警察の配備に帰している。。

  • Such neighbourhoods inherently experience a higher genuine rate of criminal activity.

    = その地域は本質的に犯罪率が高いから ━ 本文はまさにこの解釈を否定し、データは犯罪の所在ではなく取り締まりの所在を示すと論じている。。

  • The algorithms are specifically calibrated to collect more information from those areas.

    = アルゴリズムがその地域から多く情報を集めるよう調整されている ━ データ密度を生むのはアルゴリズムの設定ではなく現場の逮捕件数だと本文は説明している。。

Collocation
パターン意味
heavily patrolled厳重に巡回されたThe border remains heavily patrolled at night.
generate dataデータを生み出すEvery transaction generates usable data.
go unrecorded記録されないままになるMany minor offences go unrecorded.
under-policed district取り締まりの手薄な地区Crime festers in under-policed districts.
comparable offences同等の犯罪Comparable offences receive unequal attention.
例文
  • More patrols inevitably produce more arrests.

    巡回が増えれば必然的に逮捕も増える。

  • Unrecorded crime distorts every subsequent estimate.

    記録されない犯罪はその後のあらゆる推計を歪める。

  • The figures reflect attention, not incidence.

    その数字は発生率ではなく注目度を反映している。

  • Patrol density shapes the statistics.

    巡回の密度が統計を形作る。

💡Tip · 覚えるコツ

詳細把握では本文の因果関係を正確に追う。『密なデータ→逮捕→配備』という方向を逆転させたり、別の原因(住民・本質的犯罪率)にすり替える選択肢に注意。

3

What does the passage describe as the 'recursive validation' produced by these systems?

  • Officers sent to a flagged district make arrests there that are then fed back as confirmation of the original prediction.
  • Independent auditors review the algorithm's outputs at regular intervals to confirm them.
  • Communities voluntarily supply additional reports that corroborate the model's risk scores.
  • Analysts repeatedly retrain the model on entirely new datasets to verify its accuracy.
Phrase

fed back into the system as fresh confirmation

1級読解詳細把握
ひとことで

再帰的検証とは、配備→逮捕→システムへの再供給という自己成就のループを指す。

選択肢の解説Choice Analysis

Officers sent to a flagged district make arrests there that are then fed back as confirmation of the original prediction.

正解

本文は、モデルが地区を高リスクと判定すると警官が送られ、その存在自体がさらなる逮捕を生み、それが『fed back into the system as fresh confirmation』されると説明する。これが自己成就的予言を製造する循環であり、予測が現実を作り出すという論点を支える。

  • Analysts repeatedly retrain the model on entirely new datasets to verify its accuracy.

    = 全く新しいデータでモデルを繰り返し再訓練する ━ 本文の循環は新規データでの再訓練ではなく、同じ偏ったデータの自己強化である。。

  • Independent auditors review the algorithm's outputs at regular intervals to confirm them.

    = 独立した監査人が定期的に出力を確認する ━ 本文はむしろ democratic scrutiny からの遮断を批判しており、監査による検証は述べていない。。

  • Communities voluntarily supply additional reports that appear to corroborate the model's risk scores.

    = 地域が自発的に追加報告を提供しスコアを裏付ける ━ 裏付けは住民の自発的協力ではなく、配備された警官自身の逮捕によって生じると本文は述べる。。

Collocation
パターン意味
recursive validation再帰的な検証Recursive validation can mask underlying error.
dispatch officers警官を派遣するThe system dispatches officers automatically.
fresh confirmation新たな裏付けEach arrest counts as fresh confirmation.
self-fulfilling prophecy自己成就的予言The alert became a self-fulfilling prophecy.
purport to anticipate予期すると称するIt purports to anticipate demand.
例文
  • The prediction engineered the very outcome it forecast.

    予測はそれが予報した結果そのものを作り出した。

  • Each loop hardens the initial assumption.

    各ループが当初の前提を固定化する。

  • Confirmation arrives because attention was directed there.

    注目がそこへ向けられたがゆえに裏付けが生じる。

  • The cycle masquerades as evidence.

    その循環は証拠であるかのように装う。

💡Tip · 覚えるコツ

『recursive』『feedback loop』は、出力が入力に戻る閉じた循環を意味する。外部の新規検証(再訓練・監査)とは無関係であることを見抜く。

残り2問は、解いて確かめられます

5問すべてに、この形式の解説が付いています。 間違えた問題は記録され、あとから復習できます。

無料で登録して続きを解く

インストール不要・カード登録なし・いま全機能無料