Jun 5, 2026, 15:00

India's Top Court Makes New Rules for AI

印度最高法院為人工智慧制定新規則

Introduction

The Supreme Court of India has a new plan for Artificial Intelligence (AI). They want to use AI in courts, but people must still make the final decisions.

印度最高法院對人工智慧 (AI) 有一項新計劃。他們希望在法院中使用 AI，但最終決定仍必須由人類做出。

Main Body

AI can help with small jobs. It can find old cases, write notes, and make short summaries of papers.

AI 可以協助處理瑣碎工作。例如尋找舊案例、撰寫筆記，以及將文件製作成簡短的摘要。

AI cannot decide if a person is guilty. It cannot decide on jail time or bail. Only human judges can do these things.

AI 不能決定一個人是否有罪，也不能決定刑期或保釋問題。這些事情僅能由人類法官決定。

Lawyers must tell the court if they use AI. If the AI makes a mistake, the lawyer is responsible. The lawyer cannot say the AI was wrong as an excuse.

律師若使用 AI 必須告知法院。如果 AI 出錯，由律師承擔責任。律師不能以 AI 錯誤作為藉口。

Chief Justice Surya Kant says AI is useful. It can make the courts faster. But AI must follow the law and respect people.

首席大法官 Surya Kant 表示 AI 非常有用，能提高法院的運作速度。但 AI 必須遵守法律並尊重每個人。

Conclusion

People can give their ideas about these rules until June 20.

民眾可以在 6 月 20 日前就這些規則提出建議。

Vocabulary Learning

⚡ The Power of 'CAN' vs 'CANNOT'

In this text, we see a very important pattern for A2 learners: Ability and Permission.

1. The Positive (What is possible)

AI can help with small jobs.
It can find old cases.
It can make courts faster.

Pattern: Subject + can + action word $\rightarrow$ (Something is possible)

2. The Negative (What is forbidden/impossible)

AI cannot decide if a person is guilty.
It cannot decide on jail time.

Pattern: Subject + cannot + action word $\rightarrow$ (Something is not allowed/impossible)

💡 Pro Tip for Beginners: Notice that after can and cannot, the action word never changes.

~~AI can helps~~ $\rightarrow$ AI can help
~~AI cannot decides~~ $\rightarrow$ AI cannot decide

Vocabulary Learning

decision (n.)

A choice that you make after thinking about it

Example:The judge made a final decision about the case.

summary (n.)

A short statement that tells the main points of something

Example:Please write a short summary of the story.

guilty (adj.)

Having done something wrong or broken a law

Example:The court decided that the man was guilty of the crime.

responsible (adj.)

To be the person who must take care of something or be blamed for a mistake

Example:You are responsible for cleaning your room.

excuse (n.)

A reason you give to explain why you did something wrong

Example:I was late, but my excuse was that the bus broke down.

The Supreme Court of India Proposes New Rules for Using Artificial Intelligence in Law

印度最高法院提議制定法律領域使用人工智慧的新規範

Introduction

The Supreme Court of India has released a draft of new regulations to manage how artificial intelligence (AI) is used in the court system. The main goal is to use technology to help the courts while ensuring that human judges still make all the final legal decisions.

印度最高法院已發佈一份新監管條例草案，旨在管理法院系統如何使用人工智慧（AI）。主要目標是在利用科技協助法院的同時，確保所有最終法律決定仍由人類法官做出。

Main Body

The proposed 'Regulations for Use of Artificial Intelligence in Courts, 2026' create a clear distinction between helpful administrative tasks and forbidden judicial roles. For example, AI can be used for supportive work such as searching for past legal cases, transcribing audio, and summarizing documents. However, the rules strictly forbid using AI to decide the outcome of a case, determine a sentence, or decide if a person should be granted bail. Furthermore, the court will not allow 'black box' systems that lack transparency, nor will it allow AI to profile the behavior of legal professionals.

擬議中的《2026年法院使用人工智慧規範》在有助的行政工作與禁止的司法角色之間劃清了界限。例如，AI可用於支援性工作，如搜尋過往法律案例、音訊轉錄及文件摘要。然而，規範嚴格禁止使用AI來決定案件結果、判定刑期或決定是否准予保釋。此外，法院將不允許缺乏透明度的「黑箱」系統，也不允許AI對法律專業人士的行為進行剖析。

Accountability is a key part of this draft. The regulations state that human lawyers and judges remain responsible for any legal mistakes, even if they used AI. Consequently, legal professionals must officially declare if they used AI in their filings. If AI produces false information—a problem known as 'hallucinations'—the person who submitted the document will be held fully responsible, as technical errors are not an acceptable legal excuse.

問責制是本草案的核心。規範指出，即使使用了AI，人類律師與法官仍需對任何法律錯誤負責。因此，法律專業人士在提交文件時，必須正式聲明是否使用了AI。如果AI產生虛假資訊（即所謂的「幻覺」問題），提交文件的提交者將承擔全部責任，因為技術錯誤並非可接受的法律理由。

Additionally, Chief Justice Surya Kant has emphasized that while AI can reduce delays and improve how justice is delivered, it must follow constitutional values and respect human dignity. He asserted that technology should support the legal system but must not replace human judgment. To manage this, the court proposes creating a permanent oversight body and a specialized research center called CoRE-AI to handle the technical and ethical side of AI implementation.

此外，首席大法官 Surya Kant 強調，雖然AI能減少延遲並改善司法交付，但必須遵循憲法價值並尊重人性尊嚴。他堅稱，科技應支援法律系統，但絕不能取代人類判斷。為了管理此點，法院建議成立一個永久監督機構以及一個名為 CoRE-AI 的專業研究中心，以處理AI實施過程中的技術與倫理面向。

Conclusion

These draft regulations are currently open for public feedback. Stakeholders have until June 20 to provide their comments before the final rules are decided.

這些規範草案目前正對公眾開放徵詢意見。相關利益者可在 6 月 20 日前提供意見，隨後將決定最終規則。

Vocabulary Learning

🚀 THE LEAP: From 'Simple' to 'Sophisticated'

At the A2 level, you usually say 'Because of this' or 'And then'. To reach B2, you need Logical Connectors—words that act like bridges, showing the relationship between two complex ideas.

⚡ The 'Power-Up' Analysis

Look at these three words from the text. They don't just add information; they change the direction of the thought:

"However" $\rightarrow$ Used to introduce a contrast.
- A2 style: AI can help. But it cannot decide cases.
- B2 style: AI can be used for supportive work; however, the rules strictly forbid using AI to decide outcomes.
"Consequently" $\rightarrow$ Used to show a direct result (Stronger than 'so').
- A2 style: Lawyers use AI, so they must say it.
- B2 style: Human lawyers remain responsible; consequently, they must officially declare if they used AI.
"Furthermore" $\rightarrow$ Used to add a new, important point (Better than 'and' or 'also').
- A2 style: AI can't decide cases and it can't be a black box.
- B2 style: AI cannot decide the outcome... furthermore, the court will not allow "black box" systems.

🛠️ Practical Application: The Swap

Stop using these 'A2 words' and start using these 'B2 bridges':

Instead of...	Try this...	Why?
But	$\rightarrow$ However	It sounds more professional and academic.
So	$\rightarrow$ Consequently	It shows a logical chain of cause and effect.
Also / And	$\rightarrow$ Furthermore	It signals that you are building a stronger argument.

Coach's Tip: Use these at the beginning of a sentence followed by a comma (e.g., "Consequently, ...") to immediately signal your B2 proficiency to anyone reading your writing.

Vocabulary Learning

distinction (n.)

A clear difference or contrast between two similar things.

Example:The law makes a clear distinction between a mistake and a deliberate crime.

transcribing (v.)

The process of putting thoughts, speech, or data into written or printed form.

Example:The secretary spent the afternoon transcribing the audio recording of the meeting.

transparency (n.)

The quality of being open, honest, and easy to understand, especially regarding rules or decisions.

Example:The public demanded more transparency regarding how the government spends tax money.

accountability (n.)

The fact of being responsible for one's actions and expected to explain them.

Example:There must be greater accountability for officials who fail to protect the environment.

consequently (adv.)

As a result of something that has happened.

Example:He failed to study for the exam; consequently, he did not pass.

emphasized (v.)

Gave special importance or prominence to something in speaking or writing.

Example:The teacher emphasized that the deadline for the project was Friday.

asserted (v.)

Stated a fact or belief confidently and forcefully.

Example:The lawyer asserted that her client was innocent of all charges.

oversight (n.)

The action of overseeing or supervising a process or organization.

Example:The committee provides regulatory oversight to ensure the company follows safety laws.

implementation (n.)

The process of putting a decision, plan, or agreement into effect.

Example:The implementation of the new software took several months to complete.

stakeholders (n.)

People, groups, or organizations that have an interest or concern in something.

Example:The company held a meeting with all stakeholders to discuss the future of the project.

The Supreme Court of India Proposes Regulatory Framework for the Integration of Artificial Intelligence within Judicial Proceedings.

印度最高法院建議建立人工智慧於司法程序整合之監管框架

Introduction

The Supreme Court of India has released a draft regulatory framework intended to govern the application of artificial intelligence (AI) across all levels of the judiciary while maintaining human primacy in legal adjudication.

印度最高法院已發布一份監管框架草案，旨在管理人工智慧（AI）在司法機關所有層級的應用，同時確保人類在法律裁決中佔主導地位。

Main Body

The proposed 'Regulations for Use of Artificial Intelligence in Courts, 2026' establish a dichotomy between permissible administrative assistance and prohibited adjudicatory functions. Under the proposed regime, AI is restricted to a supportive capacity, encompassing tasks such as precedent retrieval, transcription, and document summarization. Conversely, Regulation 20 mandates an absolute prohibition on the use of AI for the independent determination of judicial outcomes, sentencing, or the assessment of bail eligibility. The framework specifically precludes the deployment of 'black box' systems—characterized by a lack of transparency—and prohibits AI-driven risk scoring or behavioral profiling of legal actors.

擬議的《2026年法院使用人工智慧條例》在允許的行政協助與禁止的裁決功能之間建立了區分。在擬議的制度下，AI 僅限於輔助能力，涵蓋前例檢索、轉錄和文件摘要等任務。相反地，第 20 條規定絕對禁止使用 AI 獨立決定司法結果、量刑或評估保釋資格。該框架明確排除部署缺乏透明度的「黑盒子」系統，並禁止對法律參與者進行 AI 驅動的風險評分或行為分析。

Institutional accountability is central to the draft, which stipulates that the responsibility for legal errors remains with the human practitioner or judicial officer, regardless of AI involvement. To ensure transparency, the regulations require legal professionals to disclose the use of AI in filings via formal declarations. Should AI-generated material be found inaccurate or fabricated, the submitting party is held fully liable, as the phenomenon of 'hallucinations' does not constitute a valid legal defense.

機構問責是該草案的核心，規定無論 AI 是否參與，法律錯誤的責任仍由人類執業者或司法官員承擔。為確保透明度，條例要求法律專業人士在提交文件時，必須透過正式聲明披露 AI 的使用情況。若發現 AI 生成的材料不準確或造假，提交方需承擔全部責任，因為「幻覺」現象不能構成有效的法律辯護。

Parallel to these regulations, Chief Justice Surya Kant has articulated a philosophical positioning on the intersection of technology and sovereignty. In a lecture at Birkbeck College, the Chief Justice characterized AI as an operational reality that necessitates a framework grounded in constitutional values and human dignity. He posited that while AI offers significant potential to enhance the administration of justice and reduce systemic delays, it must not supplant the human authorship of governing principles. This perspective aligns with the proposed creation of a permanent apex body and a Centre of Research and Excellence on Artificial Intelligence (CoRE-AI) to oversee the technical and ethical implementation of these tools.

與這些條例平行的是，首席大法官 Surya Kant 針對技術與主權的交集闡述了一種哲學定位。在 Birkbeck College 的講座中，首席大法官將 AI 定義為一種運作現實，需要一個基於憲法價值與人類尊嚴的框架。他認為，雖然 AI 在提升司法行政效率與減少系統性延遲方面具有巨大潛力，但絕不能取代人類對管理原則的主導地位。這一觀點與擬議成立的常設最高權力機構以及人工智慧研究與卓越中心（CoRE-AI）一致，旨在監督這些工具在技術與倫理上的執行。

Conclusion

The draft regulations are currently open for stakeholder consultation until June 20, pending finalization of the judicial AI governance structure.

該監管草案目前開放利害關係人諮詢至 6 月 20 日，隨後將敲定司法 AI 治理結構。

Vocabulary Learning

The Architecture of 'Institutional Weight': Nominalization and Conceptual Density

To transition from B2 to C2, a student must stop viewing grammar as a set of rules and start viewing it as a tool for conceptual density. This text is a masterclass in Nominalization—the process of turning verbs (actions) and adjectives (qualities) into nouns to create a high-register, objective, and authoritative tone.

⚡ The Pivot: From Process to Entity

Observe how the text avoids simple active phrasing. Instead of saying "The Court wants to regulate how AI is integrated," the author writes:

"...regulatory framework for the Integration of Artificial Intelligence..."

By transforming the action integrate into the noun Integration, the writer shifts the focus from the act of doing to the concept of the system. This is the hallmark of C2 academic and legal English: it treats processes as fixed entities.

🔍 Linguistic Dissection: High-Value Clusters

Look at these specific C2-level clusters and how they function:

"Human Primacy in Legal Adjudication": Here, primacy (the state of being first/most important) replaces a phrase like "making sure humans are more important than AI." The noun adjudication replaces "the act of judging."
"Dichotomy between permissible administrative assistance and prohibited adjudicatory functions": This is an extreme example of density. The author establishes a binary (dichotomy) using two complex noun phrases. Notice the adjectives permissible and prohibited acting as precise qualifiers for the nouns.
"Philosophical positioning on the intersection of technology and sovereignty": Rather than saying "he talked about how technology and sovereignty meet," the author uses positioning and intersection. This creates an intellectual distance that is expected in high-level diplomacy and law.

🛠️ C2 Application Strategy

To replicate this, stop using "action-oriented" sentences when describing complex systems.

B2 approach: We need to make sure that the people in charge are accountable for the mistakes the AI makes. C2 approach: Institutional accountability remains central, stipulating that liability for legal errors vests with the human practitioner.

The shift: Make sure $\rightarrow$ Accountability (Nominalization) Mistakes $\rightarrow$ Legal errors (Precision) Are accountable $\rightarrow$ Liability vests with (Legal Collocation)

Vocabulary Learning

adjudication (n.)

The formal act of making a judicial decision or pronouncing a judgment in a legal matter.

Example:The final adjudication of the property dispute rested with the High Court.

dichotomy (n.)

A division or contrast between two things that are represented as being entirely different.

Example:The report highlights a stark dichotomy between the theoretical benefits of AI and its practical risks.

precludes (v.)

Prevents the occurrence or existence of something; makes something impossible.

Example:The current legislation precludes the use of automated systems for sentencing prisoners.

stipulates (v.)

Specifies a requirement or a condition as part of an agreement or regulation.

Example:The contract stipulates that all disputes must be settled through arbitration.

supplant (v.)

To supersede and replace something, often by force or through strategic maneuver.

Example:Digital archives have largely supplanted physical libraries in the research process.

apex (adj.)

Referring to the highest point or the top-most level of a hierarchy.

Example:The apex court of the land has the final authority to interpret the constitution.

Practice All words in a crossword→