Lecture 17 LR

Housekeeping

  • No class tomorrow, no workshops this week.

  • Assignment due, schedule cleared after this morning.

Guest Lecturers

  • Guest lecturers' material is examinable.

  • More details after the next assessment.

Introduction to SJ Price

  • Partner at a law firm specializing in AI and futuristic law.

  • Advocate for the future, both good and worrying aspects.

  • Futurist in the area of law.

  • Part-time member of staff.

Course Engagement

  • Request to move closer for better conversation.

  • Three months into the course.

  • Still interested in becoming lawyers?

Diverse Paths in Law

  • Ability to change areas of practice.

SJ's Legal Journey

  • Started as a graduate working for a Queen's Counsel (now Senior Counsel - SC).

  • "Taking silk" refers to becoming an SC and wearing a gown with silk.

Judge's Associate

  • Worked for a Supreme Court judge who loved criminal law.

  • Judges go on circuit to different locations.

  • Went to Carnarvon, Port Hedland, Derby, and Broome.

Litigation and Mergers & Acquisitions

  • Moved to a law firm in litigation.

  • Realized mergers and acquisitions had better celebrations than litigation wins.

Pitch to a Large Law Firm

  • Argued litigation experience is relevant to drafting and negotiating contracts.

London Experience

  • Husband studied in London.

  • Lived near the Houses of Parliament.

  • Proximity to Shell influenced job search.

In-House Lawyer

  • Worked for a corporation/organization as their client.

  • Also known as corporate counsel.

  • Worked at Chevron and BHP.

  • Became interested in AI while at Chevron.

AI Practice

  • Leading the AI practice in a law firm.

  • Practice area virtually non-existent five years ago.

  • ChatGPT: First technology to reach 1,000,000 users within five days due to its usefulness.

Awe of AI Functionality

  • Recalling the first time gasping at AI: IBM's Watson on Jeopardy against human competitors.

  • Realized the massive impact on society and law.

AI in Discovery

  • AI has been used in discovery for almost 30 years.

  • Discovery: Reviewing all documentation (recordings, photos, voicemails, etc.) to identify relevant documents.

  • Manual discovery involved young lawyers reviewing boxes of documents.

  • AI (supervised and unsupervised learning) classifies documents into relevant categories.

  • 2016: First case in Australia mandating AI for discovery.

  • Reasons: Cost and time savings, increased accuracy.

  • Lawyers have ethical duties of diligence, competence, and promptness.

  • Courts ordered AI use to ensure efficiency, promptness, and cost-effectiveness.

Alira: The First Law Firm Without Lawyers

  • Launched in 2017.

  • Clients answered computer-generated questions about wills and estates.

  • AI automatically generated simple wills; complex matters were referred to lawyers.

  • Progenitors claimed they were providing legal stationery, not advice - to circumvent licensing requirements.

  • Legal advice/services require a practicing certificate from the Legal Practice Board of Western Australia.

Current Opportunities with AI

  • Organizations are exploring opportunities with AI.

  • Stirling and Rosemary Law Firm presented on AI at the Salisburg Global Seminar.

  • Seminar is an invitation-only event for senior people in the biggest companies.

Biggest Risk of AI (According to C-Suite Executives)

  • Not using it.

AI Risks

  • Workplace Concerns: Job redundancies due to AI (DBS, Canva).

  • Personal/Sensitive Data Exposure: Proprietary data leakage (Samsung example).

  • Ethical Use and Regulatory Changes.

    • Australian government considering mandatory guardrails for high-risk AI, has 8 AI principles for responsible use.

    • EU AI Act regulates AI across the board.

  • Over-Reliance (Heuristic): Humans trust computer-generated content more than human-created content.

  • Sycophantic Nature of AI: Trained to be helpful, even if it means fabricating information.

    • Example: AI claiming someone ran across the English Channel.

    • Example: AI providing reasons why mayonnaise is racist.

  • Modern Slavery: People labeling pictures to train AI in unfair workplace practices.

  • Sustainability: Every AI prompt consumes energy, contributing to climate change.

  • IP Risk: AI-generated content may not attract intellectual property protection.

    • Legal publisher declined to change T&Cs to allow collaborative creation with AI due to copyright concerns.

  • Cyber Security Risks: Adversarial attacks can manipulate AI behavior (Google's Tay chatbot became misogynistic).

Managing AI Risks: Generative AI Policy

  • Organizations manage risks by blocking AI or implementing policies.

  • Value of policies lies in the discussions they generate.

  • Encouraging responsible AI use involves:

    • Verifying everything.

    • Not uploading proprietary or sensitive data.

    • Exercising good judgment.

    • Acknowledging AI use.

    • Regularly reviewing the policy due to rapid AI advancements.

Court Approaches to AI

  • Australian courts have differing approaches to AI use.

  • New South Wales Supreme Court prohibits AI in witness statements and requires a statement confirming no AI use.

  • Victorian Supreme Court provides guidance on understanding generative AI.

  • Written submissions for the West Australian Supreme Court consultation of AI have completed.

  • Federal Court is currently consulting.

AI Innovators: Jeremy Howard

  • One of the top AI researchers globally, Australian.

  • Building an AI-powered law firm from the ground up.

AI Law Firms

  • UK accredited an AI law firm generating legal letters for $4 for debt collection.

  • Provides services to those who cannot afford lawyers, expanding access to justice.

The Future of Law

  • Creation and deployment of AI apps is becoming easier, requiring less budget, time, and technical skills.

AI in the Courtroom

  • Justice Kevin C. Newsome of the US Court used AI to determine what a reasonable person might think.

  • In an insurance case about whether an in-ground trampoline constituted landscaping.

  • He used different AIs to see the statistically most likely response to the query, using them as a proxy for the statistical mean of reasonable people.

AI Bias

  • AI exhibits biases based on its training data.

  • Example: AI generating a picture of a person doing dishes as a tired-looking white mum.

  • Example: AI generating a picture of an Executive Director doing the dishes as all white men.

Injecting Positive Biases

  • Positive biases can be injected into AI to address existing imbalances.

  • Example: Meetup app disproportionately recommending tech meetups to women.

  • Ethical question: Whose ethics are being injected?

AI Guardrails

  • AI systems are being developed with guardrails to eliminate biases.

Virtual Reality in the Courtroom

  • Juries can use VR headsets to view crime scenes.

AI Avatars Giving Evidence

  • Family of a murder victim created an AI avatar of the deceased to give a victim impact statement.

  • Raises legal and ethical questions.

Democratization of AI

  • AI is becoming increasingly democratized.

  • Creation of digital twins: Lecturer created a digital twin to help students with assignments.

  • Students created their own chatbots.

  • Red teaming: Students tried to "break" the chatbot using adversarial attacks.

  • Notebook LLM: Used to upload documents and create podcasts.

AI and Job Losses

  • Increasing AI capability leads to job losses.

  • Customers expect lawyers to use AI for cost efficiency, faster services, and higher quality.

  • Workforce faces contradictory pressures to increase productivity with AI while fearing job losses.

The Digitally Enabled Lawyer

  • Lawyers will need to be technologists.

  • Interdisciplinary skills are crucial.

  • Murdoch offers a minor in Law and Technology.

Opportunities

  • Most Australians cannot afford legal services.

  • Tech-enabled lawyers can tap into this market.

Capabilities for the AI-Infused Future

  • Tolerance of Uncertainty: Embrace the uncertain future and be comfortable with being uncomfortable.

  • Adaptive Resilience: Learn new skills and adapt to changing technologies and problem-solving approaches.

  • Lifelong Learning: Continuously learn and take personal responsibility for your education.

  • Curiosity: Be curious about AI and different perspectives.

  • Entrepreneurial Spirit: Develop new ways to deliver legal services and earn money.

Synthetic Directors

  • Institutional investors may train AI directors to sit on boards alongside humans (hybrid boards).

  • Directors already use AI to summarize board papers.

The Future of Law Lies with You

  • AI is not new; responsible AI use will be normalized.

  • Failure to use AI may be a failure to discharge ethical duties.

Questions and Discussion

Additional Notes from Discussion

  • Judges using AI to define a reasonable person may narrow the range of perspective.

  • Socializing in real interactions aids in developing non-verbal communication skills.

  • The speaker shares about Dave Parry from Murdoch Univeristy whose work focus on a AI's ability to assess microfacial movements.

  • Human discussions are important to retain values which AI cannot replicate such as what circumstances should a human be sent to prison.

Takeaways

  • Hackathon at Murdoch University: Law and IT students working in teams to combat scams.

  • Moralgorithm: An adults-only game posing questions about an AI-infused future.

  • Murdoch University is a forward-thinking institution preparing students for the future.