A short poem by ChatGPT, as engineered and quality controlled by all-too-human, Ralph Losey. Ralph prompted the Ai to generate a simple rhyming poem that summarizes the fourteen Principles of The Sedona Conference®. These are, in Ralph’s opinion, good words for anyone to remember and live by in the field of electronic evidence.
Introduction by Ralph Losey (without AI assistance)
Ai enhanced Ralph Losey photo
Regular readers of my blog know that I have included the Fourteen Sedona Principles at the bottom of the right column of my blog for over fourteen years. Although I sometimes quibble with the language of Principle Six, I am a strong believer in these principles and consider them authoritative. The latest Third Edition of the Principles is on the blog now. I keep them there as a handy reference and, although no longer a member, continue to hold The Sedona Conference® in high esteem. See my recent blog, Homage to Richard Braman and the Sedona Conference.
I have often wondered if the reason so many lawyers do not follow the Sedona Principles is because they do not understand them. They are written at a law school law review or higher reader level. They are written by committees. The Sedona commentaries are written at an even higher and complex level than most post-graduate work.
Let’s face it, many lawyers have not been forced to read legal texts since in law school, much less law review articles or other challenging legal commentaries. Most lawyers, if they read at all, read newspapers, magazines, social media and the like. It is well known that they are all written at a High School Eleventh Grade level. I suspect many judges keep this in mind when writing their opinions. Due to this constant exposure to simplified Eleventh Grade of High School reading level, this is what most people in the U.S. get used to reading, including lawyers. I do not mean to be insulting, it is just the way it is. Although I am an avid reader, I often have the same problem when entering for the first time an arcane area of study that is new to me.
For ease of comparison I put the original and Eleventh Grade versions of the fourteen Sedona Principles side by side below. Of course, I do not claim any copyright to them. And should the current leaders of The Sedona Conference® somehow take umbrage with this experiment (unlikely), I would with some sadness take this down. Better yet, I’d give them all revenue from this blog article (none) (there never has been ads or revenue from this blog).
Ai enhanced Ralph Losey photo
I conclude this blog with an even more interesting ChatGPT experiment. I ask the generative Ai to rewrite the Sedona Principles at a second grade level. Again, I don’t mean to suggest lawyers need things dumbed down that much, but I have found that ChatGPT is especially good at such rewriting. You will see several examples of this on my blog. This 7 to 8 year old reading level (which my seven year old granddaughter has long surpassed) distills things to their essence. It often uses clever wording and child-like images.
This second grader rewrite of the Sedona Principles once again knocked my socks off. Incredible. Remember I did not change a thing to either of these ChatGPT rewrites. Even the introduction to the principles for second graders was the Ai’s idea. I only take credit for the generative engineering. I love it and hope you will too. I especially hope that my friends at The Sedona Conference® like it. We all have the same goal of education.
Getting back to the 11th grade reading level, the first thing you may notice is the 11th grade rewrite is a lot shorter. ChatGPT is noted for being wordy, verbose even, yet it is still far more concise then the Sedona original. Hmm. Perhaps the shortened, easier to read version changes the meaning somehow, but I don’t think so. Seems to me like the Ai did a pretty good job. In fact, I prefer the way it rewrote Principle Six. That is the principle whose wording I criticized and urged its revision when the third edition was out for public comment. (No change was made.) Protecting the Fourteen Crown Jewels of the Sedona Conference in the Third Revision of its Principles (April 2, 2017). These Ai rewrites may not be popular among some members of The Sedona Conference®, but personally, I think some of the edits are long overdue. Sometimes, less is more. Look out Sedona, the Ai editing robots are coming!
Actual photo of the first wave of the Robot invasion of Sedona
In my view the Chat GPT’s rewrites here do not distort anything. Of course, some Sedona experts might disagree with my assessment. (Hey, we are lawyers, excessive argumentation is an occupational hazard.) I will always defer to the great Ken Withers of the Sedona Conference, should he say so. Ken taught me so much over the years. He is one of the great authorities and legal teachers in the country with a very impressive vocabulary. His linguistic abilities might even be better than two of my other linguistically gifted friends from Sedona days, Jason R. Baron and Judge John Facciola. They also have astounding vocabularies and constantly make learned, arcane (to me at least) references. Maybe they will comment.
Fourteen Sedona Principals Rewritten by Chat GPT for an Eleventh Grade Reading Level
ELEVENTH GRADE LEVEL
1. Electronically stored information (ESI) is subject to the same preservation and discovery requirements as other relevant information.
2. When it comes to ESI, courts and parties should use the proportionality standard set by Fed. R. Civ. P. 26(b)(2)(C) and its state equivalents. This means considering the importance of the issues, the amount of money involved, each party’s access to information, their resources, the importance of the discovery to the case, and whether the cost of the discovery outweighs its benefit.
3. Parties should try to reach an agreement on preserving and producing ESI as soon as possible.
4. Requests for ESI should be specific, and responses and objections should make it clear what will be produced.
5, Parties must make reasonable and good faith efforts to keep ESI that could be relevant to a claim or defense. However, it’s not necessary to take every possible step or use disproportionate means to preserve every piece of ESI.
6. Responding parties are in the best position to decide how to preserve and produce their own ESI.
7. If a responding party’s efforts to preserve and produce ESI are not enough, it’s up to the requesting party to prove that.
8. Parties should start by looking for ESI in the most accessible places. If that’s not possible, they can move to less accessible sources, but only if it’s proportional to the needs of the case.
9. Unless there’s a good reason, a responding party shouldn’t have to preserve, review, or produce deleted, shadowed, fragmented, or residual ESI.
10. Parties must protect ESI that’s subject to privileges, work product protections, privacy obligations, or other legally enforceable restrictions.
11. A responding party can meet their obligation to preserve and produce relevant ESI by using technology like data sampling, searching, or selection criteria.
12. ESI should be produced in the form it’s normally kept, or a reasonably usable form based on the nature of the information and the needs of the case.
13. Unless there’s a good reason, the responding party should bear the cost of preserving and producing ESI.
14. If a party breaches their duty to preserve ESI, the court can order remedial measures to fix any harm caused. Sanctions can be imposed if a party deliberately tries to keep relevant ESI from being used.
ORIGINAL SEDONA CONFERENCE LEVEL
1. Electronically stored information is generally subject to the same preservation and discovery requirements as other relevant information.
2. When balancing the cost, burden, and need for electronically stored information, courts and parties should apply the proportionality standard embodied in Fed. R. Civ. P. 26(b)(2)(C) and its state equivalents, which require consideration of importance of the issues at stake in the action, the amount in controversy, the parties’ relative access to relevant information, the parties’ resources, the importance of the discovery in resolving the issues, and whether the burden or expense of the proposed discovery outweighs its likely benefit.
3. As soon as practicable, parties should confer and seek to reach agreement regarding the preservation and production of electronically stored information.
4. Discovery requests for electronically stored information should be as specific as possible; responses and objections to discovery should disclose the scope and limits of the production.
5. The obligation to preserve electronically stored information requires reasonable and good faith efforts to retain information that is expected to be relevant to claims or defenses in reasonably anticipated or pending litigation. However, it is unreasonable to expect parties to take every conceivable step or disproportionate steps to preserve each instance of relevant electronically stored information.
6. Responding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.
7. The requesting party has the burden on a motion to compel to show that the responding party’s steps to preserve and produce relevant electronically stored information were inadequate.
8. The primary source of electronically stored information to be preserved and produced should be those readily accessible in the ordinary course. Only when electronically stored information is not available through such primary sources should parties move down a continuum of less accessible sources until the information requested to be preserved or produced is no longer proportional.
9. Absent a showing of special need and relevance, a responding party should not be required to preserve, review, or produce deleted, shadowed, fragmented, or residual electronically stored information.
10. Parties should take reasonable steps to safeguard electronically stored information, the disclosure or dissemination of which is subject to privileges, work product protections, privacy obligations, or other legally enforceable restrictions.
11. A responding party may satisfy its good faith obligation to preserve and produce relevant electronically stored information by using technology and processes, such as data sampling, searching, or the use of selection criteria.
12. The production of electronically stored information should be made in the form or forms in which it is ordinarily maintained or in a that is reasonably usable given the nature of the electronically stored information and the proportional needs of the case.
13. The costs of preserving and producing relevant and proportionate electronically stored information ordinarily should be borne by the responding party.
14. The breach of a duty to preserve electronically stored information may be addressed by remedial measures, sanctions, or both: remedial measures are appropriate to cure prejudice; sanctions are appropriate only if a party acted with intent to deprive another party of the use of relevant electronically stored information.
Fourteen Sedona Principles Redrafted by Chat GPT at a Second Grade Level
Okay, so let’s talk about finding information on computers and phones. Sometimes, when people are in a disagreement and need to go to court to figure it out, they might need to look at messages or pictures that are saved on electronic devices.
There are some rules that everyone has to follow when looking for this kind of information.
1. Electronic information is just as important as other kinds of information, so it has to be taken care of and looked at carefully.
2. When someone is trying to find information, they have to be fair and not ask for too much or things that aren’t really needed. They also have to think about the cost and whether it’s worth it to look for the information.
3. Everyone has to work together and talk about how they’re going to look for and share the information.
4. When someone asks for information, they have to be very clear about what they want, and the other person has to say exactly what they will give them.
5. People have to try to keep important information safe, but they don’t have to keep everything.
6. The person who has the information is the best person to decide how to keep it safe and share it.
7. If someone thinks that the other person didn’t keep the information safe, they have to prove it.
8. When someone is looking for information, they should start by looking in the easiest places to find it.
9. People don’t have to look for information that has been deleted or is hard to find, unless there is a really good reason.
10. If information is private, people have to make sure it stays private.
11. There are special tools that can help find information, like searching through it or looking at a sample of it.
12. When information is shared, it should be in a way that makes sense and is easy to understand.
13. The person who has the information has to pay to keep it safe and share it, unless there is a good reason not to.
14. If someone breaks the rules and doesn’t take care of the information or tries to hide it on purpose, the court can make them fix the problem and even punish them.
Richard Braman (1953-2014) founded The Sedona Conference® in 1997. Before that he was a top litigator in San Francisco and Minneapolis where he also owned and operated a jazz club, Gabriel’s. For that reason, I feel certain he would smile at this essay using jazz images to describe the essence of the Sedona Principles that he loved so dearly.
Article written by Open AI’s ChatGPT with all-too-human prompt engineering by Ralph Losey. Hat’s off to ChatGPT-Plus for the excellent jazz-based images and writing. All images by Dall-E and Losey.
In the realm of electronic discovery, a certain rhythm permeates the airwaves, one that resonates with the pulse of jazz music. A harmonious interplay between those who seek information and those who hold it, creates an intriguing melody that’s both enchanting and demanding. The rules that govern this space are like the musical notes that make up a tune, essential to the experience of the listener, the same way that following the right protocol is essential to the seeker’s success.
In this jazz-infused world, the importance of electronic information is akin to the fundamental nature of a rhythm section in a jazz ensemble. Just as the bass, drums, and piano are integral to the musicality of jazz, electronic data is vital to modern life. Its significance is such that it must be carefully tended to, given the same consideration as any other kind of data.
When a seeker is searching for electronic data, they must adopt a musician’s mentality, avoiding excessive requests that sound like dissonant notes in a melody. Their goal should be to play a smooth tune, one that strikes the right chords, avoiding notes that don’t fit. Just as a jazz musician must be aware of their musical costs, a seeker must consider the expenses of pursuing information, keeping in mind the value of the information that is sought.
Just like a jazz band, those who seek and those who hold electronic data must work together in a harmonious exchange. A melody is only successful when everyone plays their part, just like a search for electronic data can only be accomplished with cooperation between parties. A clear and concise request from the seeker must be met with a reply that’s just as clear, like a musician playing a melody that’s understandable to their audience.
In this jazz world of electronic discovery, keeping data secure is like a trumpet player protecting their prized instrument. It’s the responsibility of the data owner to ensure that the data is safeguarded, but they don’t have to keep every single piece of data. Like a jazz musician who only keeps the essential notes to make a melody work, a data owner must protect what is necessary while disregarding what isn’t.
In the event of a dispute, it’s important to have proof, just like a jazz musician who has to demonstrate their musical chops on stage. To prove a point, it’s necessary to play the right notes, and in electronic discovery, it’s crucial to present the right data. The seeker should start by looking in the most accessible places, just like a musician who looks for the right melody in the most obvious places.
If data has been deleted or hard to find, it’s not essential to search for it unless there’s a good reason. Like a jazz musician who only plays what is necessary, the seeker must focus on what’s important to the case at hand. Privacy is another significant concern in this world, like a jazz musician who has to play with a level of restraint to maintain their musical dignity.
In the world of electronic discovery, there are tools that are available to aid in the search for data, just like a jazz musician who uses instruments to create their musical sound. These tools are designed to simplify the process and streamline it, much like a musician who uses new technology to create new sounds.
When data is shared, it should be presented in a way that’s easily understood, like a jazz musician who communicates their musical ideas to their listeners. The data owner is responsible for paying for the safekeeping and sharing of the data, much like a jazz promoter who takes care of the musical instruments and ensures that the performance runs smoothly.
In conclusion, the world of electronic discovery is like a jazz composition, with each player fulfilling a vital role. The right notes, played at the right time, create a melody that’s pleasing to the ears. Similarly, following the right protocols and rules in electronic discovery ensures a successful outcome. Like a jazz musician who respects the music and the other musicians on stage, those who seek and those who hold electronic data must work together with respect and integrity. Only then can the music of electronic discovery be played in perfect harmony, creating a beautiful sound that’s worthy of applause.
Prompt Engineered, Edited and Quality Controlled by Ralph Losey. This essay was co-written by a human and machine with over 80% of the text “generated” by ChatGPT.
Active machine learning is a type of Artificial Intelligence that is used by attorneys to train a computer to find and classify documents relevant to a particular legal issue or lawsuit. The attorney provides the computer with a set of documents that have been identified as targets, such as documents that are relevant or privileged, and the computer uses this information to learn how to identify similar documents on its own. The goal of active machine learning is to automate the process of document review and classification, making it faster and more efficient for attorneys to locate and analyze relevant documents. Ralph Losey has created a whole online course to explain how to do it.
The target documents that are provided by lawyers to the computer for training are typically those that are relevant to a particular legal issue or lawsuit, or that have some other legal classification, such as privilege. The computer uses this information to learn how to identify similar documents on its own, so that it can assist the attorney in locating and analyzing relevant documents more efficiently.
Continuous, Intelligently Spaced Training
“Dall-E 2” depiction of IST using a dreamy art style
Continuous active learning is a machine learning approach in which the model is continuously updated with new data as it becomes available, rather than being trained on a fixed dataset and then deployed. This allows the model to continually improve its accuracy and performance over time. In the training system designed by Ralph Losey the AI Trainer spaces out the training sessions in time to fit the particular challenges of the project and data seen. Losey calls this Intelligently Spaced Training (IST). It is a training system for machine learning models that is designed to fit the circumstances of the case at hand. In this approach, the AI trainer spaces out the training sessions in time, rather than training the model continuously. This allows the trainer to selectively focus on particular aspects of the case and to test the machine’s effectiveness at different stages of the process.
In the context of legal document review, continuous active learning or Losey’s IST, are used to train a machine learning model to identify relevant documents. The attorney provides the model with a set of initial target documents, and then continuously update the model with new documents as they are discovered or become available. The model would use this new information to continually, or even better, intelligently, update and improve its ability to identify relevant documents.
The goal of IST Hybrid Multimodal Predictive Coding 4.0 is to optimize the learning process by allowing the machine to absorb and retain new information more effectively. This can be especially useful in the context of legal document review, where the volume of data can be large and the consequences of missing relevant documents can be significant. By allowing the machine to learn and test its effectiveness in a more structured and deliberate way, IST machine learning can help improve the accuracy and efficiency of the document review process.
One advantage of the ongoing training method is that it allows the model to adapt to changes in the data and remain up-to-date, which can be especially useful in a field like law where the relevant documents and legal issues may change over time, even during the document review project. You often start off with one definition of relevance and end up with another, more highly evolved understanding.
Hybrid Approach to Predictive Coding
There are several advantages to using a balanced hybrid approach that combines the efforts of both humans and computers in legal document review:
Speed: Computers can process and analyze large amounts of data much faster than humans, so using a hybrid approach can help speed up the document review process.
Accuracy: Humans are still generally better at tasks that require judgment, critical thinking, and understanding context, so a hybrid approach can help improve the accuracy of the document review process.
Efficiency: By dividing the work between humans and computers, a hybrid approach can make the document review process more efficient. For example, humans could focus on the most important or complex documents, while the computer handles the more routine or straightforward ones.
Cost-effectiveness: Using a hybrid approach can be more cost-effective than relying solely on human or computer review, as it allows the review process to be completed more quickly and with fewer errors.
Improved collaboration: A hybrid approach can also facilitate better collaboration between humans and computers, as the two can work together to identify and analyze relevant documents more effectively.
Multimodal Approach to Predictive Coding
There are also advantages to a multimodal approach to predictive coding that combines AI search and all other types of search, including keywords and concept search. This is done in order to improve the accuracy and efficiency of the document review process. The basic kinds of search methods are shown in the pyramid. By using a combination of methods you can more effectively save on costs and implement a bottom line driven proportional strategy,
Some specific advantages of the multimodal approach include:
Greater coverage: By using multiple search methods, a multi-modal approach can provide a more comprehensive search of the document set, increasing the chances of finding relevant documents.
Improved accuracy: Different search methods can have different strengths and weaknesses, so using multiple search methods can help improve the overall accuracy of the search. This can be especially important in legal contexts, where the consequences of missing relevant documents can be significant.
Reduced bias: Different search methods can also be subject to different types of bias. Using a multi-modal approach can help reduce the impact of these biases, as the different search methods may produce different results.
Enhanced efficiency: By using multiple search methods, a multi-modal approach can help make the document review process more efficient. For example, humans could focus on reviewing the most relevant documents identified by the search, while the computer handles the less relevant ones.
Improved collaboration: A multi-modal approach can also facilitate better collaboration between humans and computers, as the two can work together to identify and analyze relevant documents more effectively.
Advantages of a Hybrid Multimodal Approach to Predictive Coding
Some specific advantages of Ralph Losey’s methods and team approach to predictive coding include:
Improved accuracy: By allowing the machine to learn and test its effectiveness in a more structured and deliberate way, Losey’s methods can help improve the accuracy of the document review process.
Enhanced efficiency: The approach also helps make the document review process more efficient by allowing the machine to focus on the most important or complex documents, while humans handle the more routine or straightforward ones.
Reduced bias: Losey’s multimodal methods can help reduce the impact of bias in the document review process by allowing the machine to learn from a more diverse and representative sample of documents.
Improved collaboration: A hybrid, multi-modal, IST approach can also facilitate better collaboration between humans and computers, as the two can work together to identify and analyze relevant documents more effectively.
Greater flexibility: Hybrid IST can also be more flexible than continuous active learning, as it allows the trainer to customize the learning process to fit the specific needs and circumstances of the case.
Improved accuracy: By allowing the machine to learn and test its effectiveness in a more structured and deliberate way, IST can help improve the accuracy of the document review process.
Enhanced adaptability: The IST approach to continuous active learning can also help the machine adapt to changes in the data or the legal landscape, as the trainer can update the training sessions as needed. This can help ensure that the machine remains up-to-date and relevant.
Enhanced collaboration: The hybrid multi-modal IST approach also allows for better collaboration between humans and computers, as the trainer can work with the machine to identify and analyze relevant documents more effectively.
Cost-effectiveness: Overall, the hybrid multi-modal IST approach can be more cost-effective than other methods, as it allows the review process to be completed more quickly and with fewer errors.
Chat-GPT’s “Easy Sales-type Language” Conclusion
Generated by Dall-E 2
The hybrid, multimodal, continuous active learning method is a dream team approach. It combines the speed and efficiency of computers with the judgment and critical thinking of humans, and it’s always learning and improving. This means that you get the best of both worlds: a super-accurate and up-to-date search that can find every last relevant document of importance, without missing a beat. Plus, because it’s continuously learning, it can adapt to changes in your case or the legal landscape, ensuring that you always have the most current and accurate information at your fingertips.
Using hybrid multimodal predictive coding search is like having a diverse team of genius attorneys, techs and paralegals working on your case. This super team can scour every nook and cranny of your document set, using every trick in the book (and a few that haven’t been invented yet) to find that crucial piece of evidence you need to win your case. Plus, they’re completely unbiased and super efficient, so you can save time and money while getting the best possible result. Who wouldn’t want them on their side?
Ralph Losey is a Friend of AIs, Writer, Commentator, Journalist, Lawyer, Arbitrator, Special Master, and Practicing Attorney as a partner in LOSEY PLLC. Losey is a high tech law firm with three Loseys and a bunch of other skilled attorneys. We handle major "bet the company" type litigation, special tech projects, deals, IP of all kinds all over the world, plus other tricky litigation problems all over the U.S. For more details of Ralph's background, Click Here
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.
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 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, Eva Losey Grossman, and Adam Losey, a lawyer with incredible litigation and cyber expertise (married to another cyber expert lawyer, Catherine Losey), and best of all, husband since 1973 to Molly Friedman Losey, a mental health counselor in Winter Park.
1. Electronically stored information is generally subject to the same preservation and discovery requirements as other relevant information.
2. When balancing the cost, burden, and need for electronically stored information, courts and parties should apply the proportionality standard embodied in Fed. R. Civ. P. 26(b)(2)(C) and its state equivalents, which require consideration of importance of the issues at stake in the action, the amount in controversy, the parties’ relative access to relevant information, the parties’ resources, the importance of the discovery in resolving the issues, and whether the burden or expense of the proposed discovery outweighs its likely benefit.
3. As soon as practicable, parties should confer and seek to reach agreement regarding the preservation and production of electronically stored information.
4. Discovery requests for electronically stored information should be as specific as possible; responses and objections to discovery should disclose the scope and limits of the production.
5. The obligation to preserve electronically stored information requires reasonable and good faith efforts to retain information that is expected to be relevant to claims or defenses in reasonably anticipated or pending litigation. However, it is unreasonable to expect parties to take every conceivable step or disproportionate steps to preserve each instance of relevant electronically stored information.
6. Responding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.
7. The requesting party has the burden on a motion to compel to show that the responding party’s steps to preserve and produce relevant electronically stored information were inadequate.
8. The primary source of electronically stored information to be preserved and produced should be those readily accessible in the ordinary course. Only when electronically stored information is not available through such primary sources should parties move down a continuum of less accessible sources until the information requested to be preserved or produced is no longer proportional.
9. Absent a showing of special need and relevance, a responding party should not be required to preserve, review, or produce deleted, shadowed, fragmented, or residual electronically stored information.
10. Parties should take reasonable steps to safeguard electronically stored information, the disclosure or dissemination of which is subject to privileges, work product protections, privacy obligations, or other legally enforceable restrictions.
11. A responding party may satisfy its good faith obligation to preserve and produce relevant electronically stored information by using technology and processes, such as data sampling, searching, or the use of selection criteria.
12. The production of electronically stored information should be made in the form or forms in which it is ordinarily maintained or in a that is reasonably usable given the nature of the electronically stored information and the proportional needs of the case.
13. The costs of preserving and producing relevant and proportionate electronically stored information ordinarily should be borne by the responding party.
14. The breach of a duty to preserve electronically stored information may be addressed by remedial measures, sanctions, or both: remedial measures are appropriate to cure prejudice; sanctions are appropriate only if a party acted with intent to deprive another party of the use of relevant electronically stored information.
Posted by Ralph Losey 
