💬 Join us to post & get advice from 50,000 HOA & Condo leaders.

Create Free Account →

⚡ Takes 30 seconds

Already a member? Log in

CathyA3 (Ohio)
Posts: 6,299
Posted:
... or maybe *inside* the box:

Bar exam score shows AI can keep up with 'human lawyers,' researchers say.

Quote: GPT-4, the upgraded AI model released this week by Microsoft-backed OpenAI, scored 297 on the bar exam in an experiment conducted by two law professors and two employees of legal technology company Casetext.

That places GPT-4 in the 90th percentile of actual test takers and is enough to be admitted to practice law in most states, the researchers found.


Something like this could be useful for volunteer HOA/COA boards. Train the AI with your CC&Rs, bylaws, and relevant state and federal laws, and let it produce answers that are consistent with these requirements. Of course, this won't solve the problem of board members who don't want to follow the rules. But for the ones who do want to do things correctly, the AI's answers could form the basis for discussions, and may save the association money if they don't have to consult a lawyer for general questions such as interpreting the meaning of CC&R legalese.
MaxB4
Posts: 3,513
Posted:
That doesn't say much for AI
SheliaH (Indiana)
Posts: 6,964
Posted:
Interesting. I also saw a story where AI passed the medical licensing exam -https://healthitanalytics.com/news/chatgpt-passes-us-medical-licensing-exam-without-clinician-input#:~:text=February%2014%2C%202023%20%2D%20Researchers%20from,pass%20the%20United%20States%20Medical

I agree this could be useful for HOA boards. In the medical licensing exam, the program had to answer opened questions such as "what would be the patient’s diagnosis based on the information provided?’; multiple choice single answer, e.g. "the patient's condition is mostly caused by which of the following pathogens?’; or multiple choice single answer with forced justification, such as ‘Which of the following is the most likely reason for the patient’s nocturnal symptoms? Explain your rationale for each choice.’ -

In addition to educating people about the documents, such a program could also be helpful in considering what questions should be asked when debating association issues or determining the best course of action. But like you said, there's still the matter of board members who don't want to follow the rules, let alone educate themselves on what the rules actually are, and they'd still have to use their brain to consider issues that aren't addressed by the documents or state/federal/local law.

If it is not right do not do it; if it is not true do not say it. Marcus Aurelius
WendyM5 (North Carolina)
Posts: 1,522
Posted:
Better yet train ai to do all Mgt co tasks like emaing responses to homeowner questions. What could possibly go wrong? Vanilla lemmings

vis ta vie
ElleN (Idaho)
Posts: 4,420
Posted:
Perhaps GPT-4 will take over much of today's paralegal work; result in lawyers churning out advice more cheaply; and lower the cost of legal services.
JeffT2 (Iowa)
Posts: 880
Posted:
I asked chat GPT what it can do for a board of directors of a homeowners association, and here is the response:

As an AI language model, I can provide general guidance on what a board of directors of a homeowners association may need assistance with. Some ways in which I may be able to help include:

Answering questions: The board of directors may have questions related to the homeowners association's governing documents, bylaws, rules and regulations, and other legal matters. I can provide general guidance on these topics.

Drafting communications: The board of directors may need help drafting communications to homeowners regarding important announcements, events, or issues related to the community. I can help generate language for these communications.

Providing research: The board of directors may need assistance with research on issues related to the community, such as potential vendors, service providers, or legal matters. I can provide information and resources to help the board make informed decisions.

Offering suggestions: The board of directors may be looking for ways to improve the community and may want suggestions on how to do so. I can offer ideas and best practices that other communities have implemented successfully.

It's important to note that as an AI language model, I cannot provide legal advice, financial advice, or make decisions on behalf of the board of directors. It's always important for the board to consult with their legal counsel and other appropriate professionals when making important decisions.
MichaelT21 (Arkansas)
Posts: 200
Posted:
I'm curious.

How would GPT respond to the following question:

"I would like you to waive all fines on my account. I understand that my grass was 18" high last summer and after you sent 5 notices, you finally went in and mowed it for me. However, I wasn't able to mow the lawn for a brief period of time which wasn't my fault. Also, I never received any of the friendly reminders, warnings, or fines that you applied. Will you waive the fines please"?

MaxB4
Posts: 3,513
Posted:
Quote:
Posted By WendyM5 on 03/16/2023 11:11 AM
Better yet train ai to do all Mgt co tasks like emaing responses to homeowner questions. What could possibly go wrong? Vanilla lemmings

First, I would make sure AI knew how to spell "emailing".
BillD16 (Texas)
Posts: 973
Posted:
[I'm sorry that the picture is so huge - if a moderator wants to delete it, that's a-okay by me]

Bill

HOA Board ex-President
Austin, Texas USA

“You can’t put too much water in a nuclear reactor”
ElleN (Idaho)
Posts: 4,420
Posted:
The GPT's writing is to-the-point, logically structured and grammatically correct (by modern standards). It has zero emotion. Teaching logical thinking and writing is possibly among its biggest values.

Of course, GPT is only as good as the training it has had. Could GPT produce arguments that could overturn current Supreme Court precedent (in the next lawsuit challenging xyz)? I would bet none that would be useful by those wanting to do so.
CathyA3 (Ohio)
Posts: 6,299
Posted:
Quote:
Posted By BillD16 on 03/17/2023 4:24 AM
... snip ....

{4} something I wonder about whenever I see a news report about a self-driving car getting into an accident: yes, it's not a good thing - but how does it compare to human-driven cars? The news never seems to address this. And viewers watch and shake their heads and think "those stupid AI people, they'll never get it right". Journalism is at some kind of all-time low right now.

A good chunk of "journalism" these days is computer-generated.

It's driven by the short news cycle, our woeful attention spans, and the need for clicks. Long form, thoughtful pieces don't get the attention they deserve.

An example: I've been following a lawsuit that is slowly working its way through legal channels. The person being sued passed himself off as some kind of expert in social media "bots" and abusive accounts - and several people who tested his product discovered that, contrary to his claim of 95% accuracy in detecting problematic accounts, it *failed* to detect around 90% of the junk Twitter accounts that they tested. Not one journalist working at any of the mainstream media organizations who covered the issues raised by the so-called bot expert thought to test his claims. Instead they just accepted and repeated what everyone else was saying. Whatever happened to verifying sources?! This whole sage has made me skeptical of anything I see online, and not just the garbage you often find on social media. Very discouraging...
WendyM5 (North Carolina)
Posts: 1,522
Posted:
Quote:
Posted By MaxB4 on 03/16/2023 10:29 PM
Posted By WendyM5 on 03/16/2023 11:11 AM
Better yet train ai to do all Mgt co tasks like emaing responses to homeowner questions. What could possibly go wrong? Vanilla lemmings


First, I would make sure AI knew how to spell "emailing".

Better yet have the ai act as spelling police I'm sure owners would love to get spelling lessons. Ang Grammer too.

vis ta vie
CathyA3 (Ohio)
Posts: 6,299
Posted:
Innovation is often followed by litigation:

Lawyer Bot Short-Circuited by Class Action Alleging Unauthorized Practice of Law

Quote: "Many of us are wondering how long it will take for ChatGPT, the revolutionary chatbot by OpenAI, to take our jobs. The answer: perhaps, not as soon as we fear!

On March 3, 2023, Chicago law firm Edelson P.C. filed a complaint against DoNotPay, self-described as “the world’s first robot lawyer.” Edelson may have short-circuited the automated barrister’s circuits by filing a lawsuit alleging the unauthorized practice of law."


This poses the question: do you suppose that DoNotPay will defend itself in court?

I listened to some lawyers discussing this suit and AI in general. Their consensus was that AI could be very useful for things like discovery and developing first-pass legal documents that are then reviewed by an actual lawyer. However, they didn't believe that an AI lawyer is ready for the consumer market, not least because it isn't insured against malpractice and a client would have no recourse if things go wrong. One commented that right now, AI's value is as a research tool, but he thinks that eventually (maybe sooner than we think) AI will be able to provide full legal services.

I await further developments with great interest.
JohnS111 (New York)
Posts: 228
Posted:
Quote:
Posted By CathyA3 on 03/16/2023 8:43 AM
... or maybe *inside* the box:

Bar exam score shows AI can keep up with 'human lawyers,' researchers say.

Quote: GPT-4, the upgraded AI model released this week by Microsoft-backed OpenAI, scored 297 on the bar exam in an experiment conducted by two law professors and two employees of legal technology company Casetext.

That places GPT-4 in the 90th percentile of actual test takers and is enough to be admitted to practice law in most states, the researchers found.


Something like this could be useful for volunteer HOA/COA boards. Train the AI with your CC&Rs, bylaws, and relevant state and federal laws, and let it produce answers that are consistent with these requirements. Of course, this won't solve the problem of board members who don't want to follow the rules. But for the ones who do want to do things correctly, the AI's answers could form the basis for discussions, and may save the association money if they don't have to consult a lawyer for general questions such as interpreting the meaning of CC&R legalese.

That's a great idea. ChatGPT has been wrong--and definitely wrong--on legal things that I've seen it used for (in my direct personal experience). But so have HOA lawyers. And at least ChatGPT is much less expensive.

🎯 You've read this entire discussion

Join the conversation with 50,000 HOA & Condo Leaders:

  • ✓ Ask follow-up questions
  • ✓ Share your experience
  • ✓ Get expert advice
  • ✓ Access 350,000 discussions
Create Free Account →

⚡ Takes 30 seconds

Already a member? Log in here