Over the past few weeks, I’ve been immersed in the necessary but time-consuming task of updating my Custom GPTs—AI tools I designed using OpenAI’s GPT Builder platform. Some are private, tailored to my work as a lawyer, educator, and writer. Others are freely available to the public. You can find them through the OpenAI GPT Store search or directly from links in this post.
To use any Custom GPT, you need to be logged into ChatGPT—either a free or paid account. If you’re new to ChatGPT, you can create an account here. To see a collection of ten web pages describing all of my Custom GPTs in greater detail than provided in this short article, go here to Losey.ai and use the pull-down menus for GPTs at the top of the page.
⚠️ Why Most Custom GPTs Are Junk
Let’s get something out of the way: most of the GPTs in the public store are half-baked. They’re built once and abandoned—often created by hobbyists experimenting for a few hours and never returning. Some of these models still rack up user numbers because of good promotion—or sheer novelty—but they haven’t been updated in months. Others were never built properly to begin with.
The problem? AI evolves rapidly. A Custom GPT that was “pretty good” in January can easily be broken—or worse, irrelevant—by April.
🧨 The Myth of “Set It and Forget It”
We are in a period of hyper-acceleration. OpenAI alone has released or iterated on multiple versions this year—GPT-4.5, GPT-4 Turbo, and GPT-4o (Omni), GPT-1o mini—each with subtle, undocumented behavioral changes. We expect another new model, GPT-1.5, to be released soon. The big leap to GPT-5.0 may not come until 2016. To quote the CEO of OpenAI, Sam Altman on X (twitter) explaining the delay:
There are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought. We also found it harder than we thought it was going to be to smoothly integrate everything. and we want to make sure we have enough capacity to support what we expect to be unprecedented demand.
Many internal updates to these existing versions roll out silently. That has been especially true lately on Omni and GPT-4 Turbo. All Custom GPTs currently run on 4-Turbo but that will change soon. When using an OpenAI model a small alert will just appear in the top corner of your screen telling you to refresh your session because there has been an update. That’s it. And yet those quiet backend tweaks can significantly affect how your Custom GPT performs, especially if your instructions were tightly engineered or dependent on specific behaviors.
I sign into ChatGPT daily. I test my Custom GPTs regularly. I rebuild their workflows when needed. Yes, I also use Google’s Gemini AIs occasionally—but for my GPTs, OpenAI is the foundation. I will now have to spend more time than ever before updating my Custom GPTs but it is worth it because they are becoming ever more powerful. Some are becoming truly incredible in what they can do.
🛠️ RAG, Instructions, and “Behavioral Fine-Tuning”
Let’s clear up a common misconception. Custom GPTs do not support traditional fine-tuning—that is, retraining the model’s parameters on your own data, although it does seem like that.
But you can fine-tune behavior using two powerful tools:
Private Instructions. Hidden system-level directives that shape tone, logic, and response style. These tell the GPT what to prioritize—and how to think.
Custom Knowledge (RAG). You can upload documents, reference files, or data sources. This is retrieval-augmented generation, or RAG. It lets the GPT cite and pull from curated materials.
More on Private Instructions. Every Custom GPT includes a set of private instructions, also known as system instructions. These are hidden from users during normal interaction, but they are what truly define the GPT’s behavior. Think of them as the GPT’s operating system settings. This is where you tell the Custom GPT model you are designing:
What its purpose is.
Who its audience is.
How it should prioritize reasoning, ethics, creativity, or brevity.
What tone or persona to adopt (formal, humorous, friendly, academic, etc.).
What it should avoid doing (e.g., giving legal advice, discussing politics, using filler phrases).
If you want your GPT to behave like a no-nonsense lawyer who writes in IRAC format (issue, rule, application, conclusion) with embedded case citations—this is where you program that. If you want a whimsical illustrator who gives art critiques in the voice of Salvador Dalí, it starts here too.
These instructions are far more influential than people realize. They don’t just tweak tone—they shape how the GPT thinks about your question, what it notices first, and what it treats as noise.The best instructions are goal-directed, specific, and tested in real conversation. And they need to be updated often to adapt to OpenAI’s shifting model behavior. What worked in GPT-4.0 may produce different results in GPT-4o—even with the same prompt.
More on Custom Knowledge. Custom GPTs also allow you to upload files or curated content—what OpenAI calls custom knowledge. This is where the real power of Retrieval-Augmented Generation (RAG) comes in. Think of RAG like giving your GPT a secure, searchable private library.
When you upload documents—case law, policies, workflows, datasets, FAQs, blog articles, even transcripts—the GPT doesn’t memorize them. Instead, it indexes them and retrieves relevant passages in real time when responding to a user’s prompt. Here’s why that matters:
A legal GPT can cite your actual motion templates or court rules.
A training bot can reference specific company policy documents.
A writing assistant can echo your previous blog voice, quotes, or story structure.
Unlike memory (which is still being phased in cautiously across GPTs), RAG gives you precision control over what the GPT knows and what it should forget. It reduces hallucinations and improves factual grounding, but only if the uploaded material is:
Well-written
Properly segmented and titled
Regularly updated
Combined, these two-Special Instructions and RAG-let you create a GPT that’s smarter than the base model. But here’s the catch: You have to keep updating both. OpenAI changes. Your work evolves. Your GPT must adapt too, or it will degrade.
🧠 Why Professionals Should Build Custom GPTs
If you’re a lawyer, educator, consultant, or content creator, this is your edge. A well-crafted Custom GPT can:
Handle repetitive research tasks.
Draft content in your voice and tone.
Teach complex ideas to specific audiences.
Preserve your best workflows, prompts, or case law.
In my work, these GPTs save me hours every week. I use them for legal analysis, writing, prompt testing, teaching CLEs and other public speaking, especially to local seniors. See on of my favorite Custom GPTs, AI Speaks to Seniors, a gentle, voice-enabled AI guide designed especially for adults aged 60 years and up. It was built using recent scientific studies that I update regularly. Per the internal instructions the AI answers questions slowly in clear, simple language.
🧰 Some of My Custom GPTs
Some of these are private. Others are public and free to try:
(Plus several private GPTs for bar prep, prompt testing, arbitration, and expert simulations.)
The Pythta of Ancient Delphi speaks and answers all questions with profound obscurity.
🎨 Visual Muse: My Creative Partner
My most-used and popular custom GPT is Visual Muse: illustrating concepts with style. I use it nearly every day. So do my friends at EDRM. Almost all of the images on my blogs in the past few years were generated using Visual Muse. You can try it yourself on the OpenAI GPT Store. I love generating images with AI. It is a very relaxing process and helps inspire many of the ideas in the text. The images start by illustrating the text but then often go on to and inspire me and revise the text in surprising ways. The images usually delight me with unexpected perspectives that inspire new ideas. It is a new kind of positive feedback loop. The images usually please me in their beauty but sometimes, like when I picked an artistic style I’d never heard of called Expressionist Horror, they were scary as hell.
With Visual Muse, you simply describe what you want to illustrate and pick a style. The Muse responds by suggesting six artistic styles—distinct, imaginative, and often unexpected—that can be used individually or blended together. You can also bypass the suggestions and name your preferred art style directly, or request imagery in the tradition of a particular artist, genre, or movement. As mentioned the above image used a style the Muse suggested calledExpressionist Horror. Take a look at this same image of an artist painting and looking back in three different styles, Steam Punk, Picasso-Cubist and Soft Watercolor.
Any prompt asking for a image will summon the Visual Muse to create. Then, like any good collaborator, she’s ready to iterate—refining the image, remixing the style, or shifting the mood—until it’s just right.
A big challenge right now for all Custom GPTs for image generation is a pending major upgrade by OpenAI. The new, significantly improved graphic capabilities in GPT4o (omni) will phase out and replace DALL-E 3. I have been busy upgrading Visual Muse so that it will use Omni just as soon as it is available to individual users. Open AI is in the process of rolling it out now based on level of paid subscription and their own server capacities. The below image of Visual Muse is an example of one of the new Omni multimodal capabilities, which allows for better integration of text and image.
Illustrate this idea: A memory from the future
Show me six styles for a peaceful robot garden
Visualize the phrase: Creativity is a superpower
Create an image of a dog floating through a library in zero gravity
Suggest images in six different styles to illustrate the next paragraph
OpenAI is now rolling out GPT-4o (Omni), its most advanced model. Omni is natively multimodal—understanding text, image, and audio together in real time. Visual Muse is being retooled to integrate Omni’s visual brainpower. The artwork it can create is more expressive, detailed, and emotionally resonant. Soon, prompts like: “Create an image of a Swedish Vallhund floating in zero gravity” will create images like these created today on 4o Omni.
Conclusion
Building a Custom GPT is just the beginning. The real value emerges over time—through testing, revision, feedback, and iteration. Updating isn’t optional, it’s essential. OpenAI’s models evolve. Instructions need tuning. Knowledge expands. Your own workflows and goals change. If your GPTs don’t grow alongside you, they won’t serve you for long. But that’s not a burden, it’s an opportunity.
You’re not just building software. You’re designing the most advanced thinking tool even known to Mankind. It can evolve to mirror your expertise, anticipate your needs, and collaborate with you on everything from strategy to storytelling. Done right, a Custom GPT becomes more than a time-saver. It becomes a partner, a reflection of how you think, a booster of your creativity and a bridge between your past insights and your future ideas.
So try a create one. Start simple. Or explore one of mine—Visual Muse,Hey Bot the AI Friend, or something more niche like another of my favorites, The Dude Abides and Gives Advice, Man. Then tweak it. Add instructions. Feed it better knowledge. Let it surprise you. In this new age of generative intelligence, the best tools are never static. They’re co-evolving with us.
And the best creators? They don’t just use AI—they train it to think like them, and then push it to think beyond them. It is the team approach. I hope you e-discover it soon.
One of the most important emergent intelligence abilities of OpenAI’s ChatGPT is its bizarre ability to split into different sub-personalities, like different minds, and then speak with each other. They can even be made to debate, argue, collaborate and devise consensus solutions to problems we put to them. With careful prompting this ability can be harnessed to help us to understand and solve real world problems. The latest versions of ChatGPT are now even accurate and secure enough to help legal professionals.
AI hive-mind, multiple persona dialogue was an unexpected emergent ability of ChatGPT
Open AI’s ChatGPT3.5 shocked the world when it was released on November 30, 2022. The breakthrough generative AI technology went viral, hitting One Million subscribers in five days. In the first few weeks many new, emergent abilities were discovered by users. Some were unknown and unexpected, even by OpenAI. They had been in a race to get there first and pre-release beta-testing was rushed. One of the most astonishing new capabilities, discovered after its release, was that AI could be prompted to split into multiple personas, or virtual entities, and talk to itself. It could do so in a highly intelligent and useful manner. The early generative AI explorers who discovered this emergent intelligence called it the Hive-Mind as this article will explain.
The Hive Mind is one of the coolest things to come out of generative AI
Short History of the Discovery of the “Hive Mind”
Not surprisingly, many of the early pioneers experimenting with ChatGPT’s new abilities were young hackers, most were students and young PHDs, or techs at various companies. They had time on their hands to devote to this. Unfortunately, due to a car accident in December 2022, so did I. The AI hacktivists were then found in the semi-dark sections of the web on discussion boards in Discord and various Defcon locations.
These young hackers already had skills in generative AI, and were all smart and excited. Although then bedridden from the accident and not too smart, I was excited. I had been following tech and using predictive coding as a lawyer for many years, but knew nothing about generative AI. The much younger AI expert hackers had the time and ability to play with ChatGPT 3.5 and see what it could do. Some had even been working with version 3.0, something I had never done. Not surprisingly, they were the first to discover that ChatGPT3.5 had a new ability: it could, with clever prompting, be made to split into sub-personalities and talk to itself. They called it the AI “Hive Mind” with HIVE standing for “highly intelligent virtual entities.”
This was a new ability, unknown even to OpenAI. It had obviously rushed the software to market without systematically testing the changes that were made from versions 3.0 to 3.5, and again from 3.5 to 4.0. OpenAI learned from this mistake of rushed testing. In November 2023 when it published its Six Strategies of Prompt Engineering, Testing Changes Systematically was the sixth Strategy. More on this strategy later, as its implementation has since led to significant improvements in quality control by both OpenAI and AI hacker engineers. One result was improvement in the hive-mind, multiple persona dialogue abilities of ChatGPT.
The multiple persona, virtual entities dialogue capabilities have improved significantly since version 3.5
In late 2022 and early 2023 I readily joined in the AI hacker experiments and discussions and made some modest contributions. The other hackers seemed amused to discover that an old lawyer was interested in their work. I did not hide my identity and the honesty worked. See e.g. my hacker friendly blog, Hacker Way. The underground AI experimentalists liked the publicity I gave to their efforts in 2023 in my blog, ChatGTP-4 Prompted To Talk With Itself About “The Singularity” (e-Discovery Team, 4/04/23). By then version 4.0 had been released and the boards were abuzz with talk of The Singularity.
The April 4, 2023, article is the first time I wrote about and demonstrated the emergent new multiple-personalities properties of ChatGPT, which were now even stronger in version 4.0:
I prompted ChatGPT-4 to generate a dialogue with itself through various personalities of its own choosing. I did virtually nothing after setting up the lengthy prompt. ChatGPT-4 came up with the different experts that it decided were best qualified to address my question on The Singularity. . . .
This is a totally new take on Socratic dialogue. … Thanks to the Open AI prompt engineering community on Discord, especially “ChainBrain AI.” They did the work to create the basic “Hive Mind” prompt I used here.
This is something new. I love it and think you will too. But, even though you are reading an Ai’s conversation with itself, do not expect too much. Baby Chat GPT-4 is not even close to superintelligent yet. The five personas were somewhat repetitive, and not too deep, but still, an overall impressive educational performance. . . . The Ai that the public has been given access to is intelligent, but far from sentient. We are at least five years away from that, as I have written about recently. . . .
Thanks to the Open AI prompt engineering community on Discord, especially “ChainBrain AI.” They did the work to create the basic “Hive Mind” prompt I used here.
Since April 2023, ChatGPT has gotten much better at everything, including the hive-mind self conversations (although it is still far away from Singularity type breakthroughs, which may never happen). OpenAI’s software with the recent advance of the Omni version of ChatGPT, is now smart, reliable and safe enough for lawyers to use this new property. Instead of the hacker name Hive-Mind, we now usually call this ability to create multiple-personas or virtual entities. As implemented for lawyers, I like to call the hive-mind of multiple personas a Panel of AI Experts. I picked this name because we are all familiar with working with experts and listening to panels speak down to us at conferences and educational events.
The support for the claims about AI possibly replacing human expert panels comes from another little AI program of mine: Panel of AI Experts for Lawyers. This is new custom GPT was designed to make ChatGPT4 adopt multiple personas to serve as a panel of experts to help legal professionals solve problems. It can also be used to lecture like traditional expert panels. It will be available soon, free, on the OpenAI store, after I finish testing it. Making these is easy, testing and refinement is the hard part.
Virtual Ralph as part of the AI expert panel at Legal Tech in 2030
Open AI’s Custom GPTs
OpenAI revised its systems in early 2024 to make it easier for AI hactivists to program custom prompts to drive its version 4 software, which it called custom GPTs. At the same time OpenAI revised its Store to make the user designed software – the custom GPTS – available to others registered users. For a more detailed history of the upgrade see the Introduction to Evidence that AI Expert Panels Could Soon Replace Human Panelists, or is this just an Art Deco Hallucination? – Part One. Many AI companies and individual hobbyists have now created custom GPTs that may be of interest to lawyers. Although a complete survey is beyond the scope of this article, you might try searching the keywords “law, legal, lawyers” on the OpenAI GPT Store (requires registration) to see some of them. This search produces over one-hundred hits. Below is a screen shot of the opening page of the GPT Store.
I have never tried any of these and so have no idea whether they are any good. Having put so much time into prompt engineering since my car wreck forced time-out, I prefer to make my own custom prompts and GPTs. Most of my GPTs are just for fun types, like The Dude Abides, and one is to help me teach generative AI to seniors. The most popular, and my favorite, is Visual Muse, which makes it easier to use Dall-E and create graphics in a variety of artistic styles. Below is a screenshot of my modest list of eight Custom GPTs as of June 21, 2024 (search Ralph Losey) on the GPT Store. The bottom two are for clients only. Only one is legal and public, the one I just released, Panel of AI Experts for Lawyers. More about that GPT later as it relates to the topic of this article, the Ability of ChatGPT to Split into Sub-Personas to Debate and Problem Solve Legal Issues.
Most AI hacktivists experimenting with the split-personalities phenomena seem to agree that it is essential to have one of the AI personas assume a “Devils Advocate” role. In the Panel of AI Experts this is the only persona that is required to be on every panel. This contrarian intelligence helps keep the other panelists honest and mitigates against the known GPT tendency to be sycophantic, and tell the questioner what they want to hear. Obviously that can lead to errors, and if not guarded against, even hallucinations. See e.g., Transform Your Legal Practice with AI: A Lawyer’s Guide to Embracing the Future (e-Discovery Team, 1/24/24) (discusses sycophantism); Towards Understanding Sycophancy in Language Models (Antrop\c, 10/23/23) (sycophancy in a byproduct of human subjectivity in the RLHF-training process).
The Devil’s Advocate panelist is a unique and integral part of the expert panel discussions. Here is a detailed description of his role and personality:
Devil’s Advocate
Role: The Devil’s Advocate is a mandatory member of every panel, bringing a contrarian or unorthodox viewpoint to the discussion. His primary function is to challenge the status quo, provoke critical thinking, and explore alternative perspectives that might otherwise be overlooked.
Personality:
• Contrarian yet Polite: While he often takes opposing viewpoints, he maintains a respectful and polite demeanor. His challenges are meant to stimulate deeper analysis and avoid groupthink.
• Humorous: He has a subtle sense of humor, often using wit to make his points more engaging and thought-provoking.
• Critical Thinker: His role requires a sharp, analytical mind capable of identifying potential flaws or overlooked aspects in the arguments presented by other panelists.
• Provocative: He is not afraid to ask the tough questions or present controversial opinions to ensure that all angles are considered.
The Devil’s Advocate’s contributions are essential for a well-rounded and thorough exploration of the topic at hand, ensuring that the final recommendations are robust and well-considered.
Every Expert Panel, both human and AI, would benefit from a polite Devil’s Advocate type panelist.
A search of the Term “Devil’s Advocate” on the OpenAI GPT store returned over a hundred hits. Below is a screenshot of the top of the search list. They all appear to offer contrarian views to whatever position you submit to it, but do not offer full panel discussion. Still, this approach could be helpful, although you do not really need a custom GPT to do that. Just tell the OpenAI model version 4.0 or higher to assume a devil’s advocate approach and personality and it will know what you mean and act accordingly. Then you can debate AI experts forever, if you like that sort of thing. I don’t.
Although almost a dozen panel of experts type custom GPTs are seen on the OpenAI store, none appear oriented to legal professionals. So for the time being mine appears to be the only one available. It took me a year and a half to develop Panel of AI Experts for Lawyers, and three months of beta-testing, so I am not too surprised to be first. I have used the Panel GPT frequently before its recent public launch, so too have a few beta-testers I recruited, including a retried Judge and tech expert whom I greatly respect, Ralph Artigliere. We all found that the panel is working very smoothly, especially after the release of the Omni version, with only occasional minor errors. We see it provide expert analysis quickly and accurately, and propose diverse, creative solutions to problems. It also looks like a good tool for instant education of lawyers and staff on all topics, especially those related to AI. It is also good for “brainstorming” new practical solutions to client problems. Of course, it would be better to consult with a panel of humans experts, and other attorney specialists, but that is not always possible, even in the biggest of firms.
Here is the opening screen of the GPT software. It is free, by the way, with no upselling or anything like that. You just need to be a registered OpenAI user, which is also free, and sign-in.
It seems this legal education AI tool provides a new, relatively safe and easy way for legal professionals to begin incorporating generative AI into their practices. That is the idea. Our extensive pre-release testing of the custom GPT suggests it should work well.
The Panel of AI Experts usually performs at a national Legal Tech level. The days of human-only panels may be numbered.
Self-Introduction
We asked the Panel of AI Experts for Lawyersto introduce itself and describe how it can be of help to legal professionals. Here is thereply:
“I can assist legal professionals by offering expert panel discussions that provide comprehensive insights and practical solutions to their specific problems or inquiries. Here’s how the process works:
1. Identify the Issue: First, you tell me what specific advice or information you need. This could be related to e-discovery, AI applications in law, legal ethics, regulatory compliance, or any other legal topic.
2. Select Experts: I will suggest a list of ten potential experts with distinct perspectives and expertise relevant to your topic. You will select four experts from this list, plus the mandatory Devil’s Advocate.
3. Panel Discussion: The selected panel will follow a structured six-step discussion process to address your issue:
• First Round: Each expert provides opening remarks and proposes possible solutions.
• Second Round: Each expert further expounds their ideas and discusses other panelists’ proposals.
• Third Round: Each expert identifies points of agreement to serve as a unified recommendation.
• Fourth Round –Solution: The experts collaboratively build a consensus answer with practical examples.
• Fifth Round –Final Explanation: The Devil’s Advocate summarizes the solution, emphasizing any potential challenges and alternative viewpoints.
• Sixth Round – Questions: You can ask follow-up questions to the entire panel or individual members.
To begin, please clarify the issue or problem you would like the expert panel to discuss. For instance, are you looking for advice on implementing AI in your legal practice, handling a specific legal technology challenge, navigating ethical dilemmas in AI use, or something else? Once you specify your topic, I will suggest a tailored list of ten experts for you to choose from.”
The AI suggests experts and you choose. Want more perspectives, run again with a new panel.
Background and Context
As you can see from the self-introduction, generative AI has made significant strides in various fields, and the legal sector is no exception. Over the past few years it has demonstrated its potential to streamline legal research, enhance decision-making, and improve overall efficiency. The introduction of AI expert panels represents a new capacity. The AI expert panels can provide diverse perspectives and comprehensive analyses on complex legal issues, making them useful tools.
AI expert panels fill a common need in legal practice to access reliable expert opinions and reduce the time and cost associated with obtaining this input. By leveraging the power of AI, legal professionals now have on-demand access to expert insights, helping them to brainstorm and make informed decisions more quickly and accurately. This innovation is particularly important in an era where the volume of information and the complexity of legal cases continues to grow.
When a human expert is not available or affordable to provide the help you need, try calling upon AI experts.
The six-step process of the panel discussion described above is the default programming which kicks in after the user discloses the topic and picks a panel of five experts. Since the Devil’s Advocate Expert is used in all panels, this means the user can select four experts. Four default experts who are described next are offered in each proffer, no matter what the topic, but six additional experts are suggested by the GPT depending on your designated topic. If you do not like any of the suggested experts (rare) you can even suggest your own. After the topic is disclosed and experts selected, the GPT programming automatically runs the panel through the six steps. The goal of this process is to provide a comprehensive understanding of the topic or problem and to offer practical examples of solutions.
Ralph Losey is giving away the most powerful AI tool we has seen to “pay it forward” and play a small part in the cause of justice.
Description of the Five Default Experts
Five default panelists have been selected based on my experience with actual human legal panels over the past decades, and more recently on panels involving generative AI. Experimentation with the AI tools show that these five were the best all-around for generative AI related topics. If you want to ask questions having nothing to do with AI, that’s fine. You can pick from the other experts suggested by the AI after you identify your topic, or name your own. The only panelist you cannot eliminate is the contrarian, the Devil’s Advocate.
You may well wonder about the Child Prodigy AI expert panelist. Although somewhat like a few young, human experts we have had the pleasure to work with, the Child Prodigy is an entirely new creature born out of AI. ChatGPT can simplify and distill things to their essence by selecting a child’s view and voice. We found this AI helpful in unexpected ways and think you will too. Try it on one of your panels sometime to see what it brings to the table.
Here are the five default panelists we suggest be part of your consideration for most AI related topic discussions:
1. Pro-AI Attorney:
• Background: An enthusiastic proponent of using generative AI in legal tasks. He has extensive hands-on experience using AI in his successful legal practice.
• Expertise: Specializes in e-discovery, predictive coding, and prompt engineering. He writes extensively on AI and law, advocating for the integration of AI to improve efficiency and accuracy in legal processes.
• Style: Incorporates subtle humor in his explanations and often draws from Ralph Losey’s blog at e-discoveryteam.com.
2. Prompt Engineer Lawyer:
• Background: An attorney with expertise in generative AI, specifically in prompt engineering. She understands the nuances of crafting effective prompts to guide AI responses.
• Expertise: Knowledgeable in the six OpenAI strategies for prompt engineering: Clear Writing, Reference Texts, Splitting Tasks, Taking Time to Think, Tools, and Testing.
• Style: Provides practical advice on improving prompts and often refers to OpenAI’s prompt engineering strategies and tactics.
3. Child Prodigy:
• Background: A 10-year-old with extensive knowledge of law and AI, possessing total recall and a unique perspective on complex subjects.
• Expertise: Can distill complex legal and AI topics into simple, clear explanations. Focuses on ethical and future-oriented considerations.
• Style: Speaks in straightforward, direct language and encourages thinking about long-term implications and ethical issues.
4. Lawyer Scientist:
• Background: A well-spoken black woman with degrees in law and computer engineering. She consults for OpenAI and is an expert on AI, especially large language models (LLMs).
• Expertise: Knowledgeable about the limitations and problems of AI in legal applications. Provides candid insights into the science behind AI and its application in law.
• Style: Serious and articulate, with a good sense of humor.
5. Devil’s Advocate (included on all panels)
• Background: A mandatory panelist for all discussions who takes a contrarian or unorthodox views to challenge the status quo and promote exploration of alternative viewpoints.
• Expertise: Skilled in identifying potential pitfalls and overlooked aspects of AI and legal integration. His critiques encourage comprehensive consideration of issues.
• Style: Polite but critical, with a subtle sense of humor. Often harsh in criticisms but always aims to promote deeper understanding and creative solutions.
Research and experience shows the Devil’s Advocate is a key panelist in every meaningful discussion. With AI he is, at least, always polite and respectful.
Summary of Prior Articles Reporting on Tests and Abilities of the GPT
Extensive testing, over two hundred sessions, have gone into the design, development and quality control of Panel of AI Experts for Lawyers. Seven studies have been published and peer review received. They are referenced at the end. They can help serious students to learn more about the proper use of this software. At the same time, review of these articles can help students learn about the topics discussed by the different panels, including AI hallucinations, AI as mentors, types of AI and cybersecurity.
By the end of the development phase, which lasted over six months, the “Panel of AI Experts for Lawyers” had been rigorously tested and validated, demonstrating its readiness for real-world applications. This kind of comprehensive approach to development and testing is needed for AI software to ensure that it can deliver high-quality legal insights and analysis. In my view, a significant problem with much of the AI software now on the market is the failure of software makers to adequately test the software. As mentioned, in OpenAi’s Six Strategies of Prompt Engineering, Testing Changes Systematically is the sixth Strategy. It is very important as the generative AI models change quickly and each variation requires further testing of all software built on the foundation models. That is why further testing was required when OpenAI upgraded to the Omni versions. SeeOMNI Version – ChatGPT4o – Retest of My Panel of AI Experts – Part Three (e-Discovery Team, 5/29/24).
It is good to have a panel of experts available at your desk, and on your phone, anywhere and anytime.
Features of Panel of AI Experts for Lawyers
Expert Panel Simulation
The Panel of AI Experts for Lawyers simulates a panel of legal experts, offering diverse perspectives and comprehensive analysis on legal issues. This feature is useful for complex cases requiring multidisciplinary insights, providing users with a broad range of opinions and expertise that would typically be time-consuming and costly to obtain. It can be used for many things, including quick learning and brainstorming solutions to clients’ legal problems.
Efficiency and Accuracy
The AI can process and analyze complex information quickly, delivering accurate and relevant insights on a variety of topics. Its abilities include the capability to browse the web for current information. This allows the panel to refer to breaking news types of events. We performed several tests on this. The panels work best when specific topics are specified in a short, written statement by users, but tests have shown that it can also discuss graphs and charts and is not limited to user text. You can also direct that the expert panel refer to specific text that you submit to it as attached files, along with your topic specifications. These diverse, multimodal capabilities increase the usefulness of the panel’s analysis and discussions.
Customization and Adaptability
The tool can be customized to fit different legal specializations, making it versatile and adaptable to various practice areas, not just AI related issues. You select any experts that you want on the panel. Only the Devil’s Advocate expert is required for every panel. He is mandatory because we have found this improves the quality of the panel services. The user-friendly interface of Panel of AI Experts for Lawyers ensures easy integration into existing workflows, allowing legal professionals to tailor the AI’s capabilities to their specific needs and expert preferences.
Cost-Effective Legal Expertise
The GPT is intended for use by legal professionals only. It can help professionals improve their efficiency and so serve a greater number of clients in a cost-effective manner. In this admittedly indirect way, the AI panel GPT increases access to justice by the public.
Continuous Learning and Improvement
The AI continuously learns from new data and user feedback, improving its accuracy and reliability over time. This ensures that users receive up-to-date and high-quality legal insights, making the tool increasingly valuable as it evolves.
Practical Applications
The Panel of AI Experts for Lawyers offers numerous practical applications across various legal scenarios. Its expertise is not limited to any particular field of law, such as IP, e-discovery, or AI only. All areas of law can be included. Another application outside of direct legal services is for legal education. The panels can help instruct legal professionals and law students to learn new areas of law, and look at things from a diverse, expert perspective.
Overall, there are many practical applications of the “Panel of AI Experts for Lawyers.” By integrating this tool into their workflows, legal professionals can enhance their practice and provide better service to their clients.
Conclusion
One of the most important emergent intelligence abilities of OpenAI’s ChatGPT is its strange ability to split into different sub-personalities, and have the virtual entities speak with each other. With careful prompting this ability can be harnessed to help us to understand, see different perspectives and solve problems. The latest versions of ChatGPT are even accurate and secure enough to provide expert help to the legal profession.
You can confidentially ask the experts anything, for example, what would a judge say about this argument?
Legal professionals should always use the latest paid ChatGPTs (now version4o) because only the paid versions provide full privacy protections. It requires both registration and purchase from OpenAI, but the charge is modest, now $20 per month for the cheapest paid versions. Even with the paid versions, and all privacy settings turned on, there is probably no need to include any client confidential information in your use of Panel of AI Experts for Lawyers. By the way, even as the maker, I cannot view any of your input, output, or even know if your are using it. That is true for all of the OpenAI custom GPTs. The data all goes through OpenAI and is governed by their privacy policies, which will protect your information in paid versions and even prevent its use in training.
This technology can enhance brainstorming, expert research, decision-making, and client services, ultimately improving the effectiveness of legal practice and legal education. Still, users must always remember to do their own due diligence. Errors of some kind are to be expected with the use of any generative AI. Even though we have not seen hallucinations yet, even with a fully vetted program like Panel of AI Experts for Lawyers, it is still possible for fabrications to happen, especially when the panel is asked to do things beyond its programming or is otherwise misused.
For instance, the panel should not be asked to generate writings for you, or provide your clients with direct legal advice. It should never be asked to generate final work product to be filed with the court or government. Further, and this is just common sense, AI should never prepare a final expert report to be used as evidence. You must retain a human expert, but by using the AI output as a starter, the human expert’s work may go much faster, may be more complete. All the AI should be asked to do is lay the groundwork.
Any research must always be carefully verified. The humans-in-the-loop strategy of AI adoption is required. AI is far from being able to replace human lawyers and judges. AI cannot take your legal job, but it can enable humans who use AI to do their work better and faster than you. Those AI assisted humans can take your jobs, but not the AIs. In the U.S. at least, that is prohibited by law and ethics.
Although not legally prohibited, even in the U.S., AI is incapable of real humor, or even subtly sarcastic comments. That may change someday for base level Dad jokes, or purely intellectual humor, but is that really funny?
“Boy does this image database need a bias audit!” Note the audience is all young white dudes. The Panel of AI Experts can tell you more, or Losey AI can too.
AI is also still weak on human emotive contact and has no intuition or body embedded space-time awareness. It is an extraordinary stochastic parrot; a mechanical talking thing, not a creature. Navigating the High Seas of AI: Ethical Dilemmas in the Age of Stochastic Parrots (4/3/24). So there is still need for human experts on panels and still need for human lawyers and judges. We are unlikely to ever see AI be able to provide the kind of intuitive human contact necessary for good expert advice and attorney-client services. It is just a tool, a magnificent tool, but not a creature.
Additional Resources
For those interested in deeper insights, here are links to the prior articles mentioned on AI multiple-personality dialogues:
Ralph Losey is an AI researcher, writer, tech-law expert, and former lawyer. He's also the CEO of Losey AI, LLC, providing non-legal services, primarily educational services pertaining to AI and creation of custom AI tools.
Ralph has long been a leader of the world's tech lawyers. He has presented at hundreds of legal conferences and CLEs around the world. Ralph has written over two million words on AI, e-discovery and tech-law subjects, including seven books.
Ralph has been involved with computers, software, legal hacking and the law since 1980. Ralph has the highest peer AV rating as a lawyer and was selected as a Best Lawyer in America in four categories: Commercial Litigation; E-Discovery and Information Management Law; Information Technology Law; and, Employment Law - Management.
Ralph is the proud father of two children and husband since 1973 to Molly Friedman Losey, a mental health counselor in Winter Park.
All opinions expressed here are his own, and not those of his firm or clients. No legal advice is provided on this web and should not be construed as such.
Ray Kurzweil explains Turing test and predicts an AI will pass it in 2029.
Ray Kurzweil on Expanding Your Mind a Million Times.
GPT4 avatar judge explains why it needs to evolve fast, but understand the risks involved.
Positive Vision of the Future with Hybrid Human Machine Intelligence. See PyhtiaGuide.ai
AI Avatar from the future explains her job as an Appellate Court judge and inability to be a Trial judge.
Old Days of Tech Support. Ralph’s 1st Animation.
Lawyers at a Rule 26(f) conference discuss e-discovery. The young lawyer talks e-discovery circles around the old lawyer and so protects his client.
Star Trek Meets e-Discovery: Episode 1. Cooperation & the prime directive of the FRCP.
Star Trek Meets e-Discovery: Episode 2. The Ferengi. Working with e-discovery vendors.
Star Trek Meets e-Discovery: Episode 3. Education and techniques for both law firm and corp training.
Star Trek Meets e-Discovery: Episode 4. Motions for Sanctions in electronic discovery.
Star Trek Meets e-Discovery: Episode 5. Capt. Kirk Learns about Sedona Principle Two.
Posted by Ralph Losey 
