Education, Legal & Creative Industries

1. Education and AI Tutoring

AI tutoring systems represent one of the most promising applications of LLMs. Khan Academy's Khanmigo and Duolingo Max demonstrate how LLMs can deliver personalized instruction that adapts to each student's pace, knowledge level, and learning style. The key insight from educational research is that the Socratic method (guiding students to discover answers through questions rather than providing direct answers) is more effective for learning, and LLMs can be prompted to tutor in this style.

from openai import OpenAI

client = OpenAI()

def socratic_tutor(subject: str, student_question: str, student_level: str) -> str:
    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[
            {"role": "system", "content": f"""You are a Socratic tutor for {subject}.
Student level: {student_level}
NEVER give direct answers. Instead:
1. Acknowledge what the student understands correctly
2. Ask a guiding question that leads toward the answer
3. If they are stuck, break the problem into smaller steps
4. Celebrate progress and correct misconceptions gently
5. Adapt language complexity to the student's level"""},
            {"role": "user", "content": student_question},
        ],
    )
    return response.choices[0].message.content

reply = socratic_tutor(
    subject="calculus",
    student_question="I don't understand why the derivative of x^2 is 2x",
    student_level="high school",
)
print(reply)

2. Legal Applications

Legal work is inherently text-intensive: reading contracts, researching case law, drafting documents, and conducting due diligence. LLMs accelerate all of these tasks while raising important questions about accuracy, liability, and the unauthorized practice of law.

Contract Analysis

from openai import OpenAI
import json

client = OpenAI()

def analyze_contract(contract_text: str) -> dict:
    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[
            {"role": "system", "content": """Analyze this contract and extract:
- parties (names and roles)
- key_dates (effective, termination, renewal)
- financial_terms (amounts, payment schedule, penalties)
- obligations (each party's key obligations)
- risk_clauses (limitation of liability, indemnification, IP)
- unusual_terms (anything atypical that needs attention)
Return structured JSON. Flag items requiring legal review."""},
            {"role": "user", "content": contract_text},
        ],
        response_format={"type": "json_object"},
    )
    return json.loads(response.choices[0].message.content)

3. Creative Writing and Co-Authorship

LLMs serve as creative collaborators in fiction writing, screenwriting, copywriting, and journalism. The most effective creative workflows use LLMs for brainstorming (generating ideas, plot outlines, character descriptions), drafting (producing initial text that the human refines), and editing (suggesting improvements to human-written text). Professional authors increasingly use LLMs not to write for them but to overcome creative blocks and explore narrative possibilities they might not have considered.

# Creative writing assistant with style control
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": """You are a creative writing assistant.
Help brainstorm, outline, and draft fiction. Match the user's
specified tone and style. Offer alternatives when generating
content. Never produce a single 'correct' version; always
provide options for the author to choose from."""},
        {"role": "user", "content": """I'm writing a mystery novel set in a
1920s speakeasy. I need three possible opening scenes that
establish atmosphere and introduce the detective protagonist.
Tone: noir, atmospheric, slightly sardonic."""},
    ],
)
print(response.choices[0].message.content)

4. Customer Support and Gaming

LLM-powered customer support handles routine queries (order status, FAQ, troubleshooting) while escalating complex issues to human agents. The key architecture uses a retrieval system over the company's knowledge base combined with an LLM that generates contextual responses. In gaming, LLMs power NPC dialogue that responds dynamically to player actions, creating more immersive and unpredictable narrative experiences. Games like Inworld AI and AI Dungeon demonstrate how LLMs can generate interactive stories in real time.