OG Talks: Good Energy – AI’s Impact on Oil & Gas in 2025

Hello, and thank you for joining us for this edition of OG Talks Good Energy. I’m here today with my partner, Molly Pila, to talk a little bit about the, effect that we see AI having on the future of litigation. Molly, thank you for joining me.

Thanks for having me.

AI is everywhere these days. You can find what’s good, what’s bad, horror stories, sanctions, all kinds of different uses, and a lot of, a lot of different people who have a lot of different perspectives on where things are going.

Today, I wanna talk a little bit about where you see things headed in the litigation realm and in particular, why that might be of interest to our oil and gas producer clients.

So what are some of the challenges and limitations, that you’re seeing these days with AI?

Well, the limitations, you know, stop me when you’ve heard this oil and gas, little conservative of an industry. And so some folks are really kinda reticent to adopt AI, just because all this information goes in. It goes into this magical black box, and then input comes out. And so there’s a hesitancy just to adopt it given the complexities of oil and gas.

And so it’s like, well, did you consider this? Have you considered this lease line? Have you considered, you know, this depth? Have you considered, you know, is all of this baked into the mix?

And and not being able to really see kind of what the pathway is of the algorithm, it’s difficult to believe, like, I can trust this material. Right? I can I can rely on this, and this is good? And, you know, particularly because you’re dealing with folks that are investing millions and millions and millions of dollars into this, you know, to believe like, oh, that’s what the computer told me, is is a difficult sell sometimes when people are trying to make a decision about spending millions of dollars.

Yeah. That makes a lot of sense. You know, one of the phrases I’ve heard when it comes to AI is garbage in, garbage out. And I think one of the concerns in our particular industry, especially when it comes to land rights, mineral rights, is that every parcel of land is unique. And every parcel of land has its own issues, its own ownership, its own nuances.

And if you’re not feeding the exact correct data into AI, then what you get out is probably not going to be accurate. And as it stands right now, some of these services like chat, GPT, and others, they don’t necessarily have an untethered, access to every courthouse’s records, every county’s records, every single document that might be important. So do you think it’s fair to say that right now, there are these types of limitations?

For sure.

And, you know, it’s interesting that you talk about chat GPT too because another one of the drawbacks is the energy that it takes to funnel chat GPT versus just like a Google search or a normal search just kind of on the Internet.

It’s like a nine tie it’s like a nine x difference. Right? So it’s like two point seven watt hours of electricity for a chat GPT search versus point three for Google. So there is a, you know, what are we saving here by using chat GPT relative to this? And, yes, absolutely to your point is there there’s so many records that are recorded and so many records that aren’t recorded that all go into making sure that you’ve got the proper information in order to, you know, predict where it is you should drill and who you should pay on the back end if, God forbid, and as we’re all hoping to do, actually get production from from these spaces.

In AI, they’re just gaps. And, you know, another weakness that AI has a little bit is, you know, this this talent vacuum. You take a highly specialized field like oil and gas and a highly specialized field like AI. You’re trying to combine it. It’s like trying to put two unicorns together a little bit and come up with a product that’s reliable that you can look at and say, I believe that this information is correct and I can trust this.

Yeah. So I think you’re still gonna need a a human touch and and human oversight. And, I think that our clients out there probably wanna be cautious of somebody who’s really pushing AI as a big cost cost saving measure right now because they’re really probably cutting other corners to do so, and they’re and they’re doing that without kind of that proper oversight, you know, and somebody with a critical eye really looking at the results, looking at the source material, making sure that that AI is getting the complete picture. And I think that we’re probably still some time away from that. I mean, would you kind of agree with that?

Yeah. Look. God bless the Internet. You can find anything you want on the Internet relative to AI.

You can find all of this information that says the robots are coming and are going to take over the world and humans will become obsolete. And then you’ll find other folks that say AI is gonna be the, you know, like, next laserdisc era of, you know, a flash in the pan, and then it’ll be, you know, here and then gone. I’m gonna take the neutral approach because I think it’s the real one is that AI is gonna be an incredible tool, eventually, and once it’s built out and once there is some more reliability to it. But you’re always going to need a human to look at it and oversee it and finally make give these decisions about, you know, is this where we’re gonna drill?

Is this the time to do it? What do these production numbers need to look like? You know, how do we take this information and actually put it together and and assess it? And look.

AI’s got a lot of incredible I mean, they’re they’re making tremendous, tremendous amounts of advancement. But, you know, speaking about cost measures as well, you’re you’re looking to integrate this super powerful, super, you know, time demanding AI, you know, and then integrate it with software that’s not at those speeds and everything. And so either you gotta switch whole cloth to the new AI system, which people are wildly uncomfortable doing, or try to integrate it with the systems that work, which is slow and time consuming and expensive.

Well, I still have my Star Wars laser discs, and I’m convinced that they’re going to make a comeback. So I’ll just put that out there.

Is it on eBay?

EBay.

They’re they’re they’re, you know, they’re maintaining their value. I’ll put it that way.

So let’s talk about some of the ways that that AI is being used and is being helpful.

One of the things that we’re seeing is AI type solutions being used to analyze a lot of geological data, reservoir analysis, looking at results, looking at mud logs, helping to optimize location scouting, things like that.

Are you seeing clients, kind of moving that direction?

Sure. I mean, it’s it’s great. Right? If you can use this information and get a more accurate prediction of these, you know, subsurface reservoirs and where you’re I mean, more likely to hit. You know, you’re preventing waste from a money perspective. You’re preventing waste from, you know, potential environmental implications, whether, you you know, drilling dry holes or, you know, just from, again, the energy it takes to drill it.

You know, yeah, folks are using it again as a tool. Smart, smart folks are using it as a tool, and then ultimately still having a decision maker on the back end. And it’s really helping with, you know, forecasting, drilling schedules, and managing production. And, you know, when you think of the five hundred thousand things that you’ve gotta put into, you know, producing and getting a well ready of just from a hiring perspective, from an equipment perspective, from, you know, just all of the logistics and insurance and, you know, mobilization and all of those things. It can be super helpful to predict those, you know, where to drill and when to drill. And, you know, it tracks the market and can give you real time indications as to cost and production and whether or not that’s valuable. And it’s it’s huge.

That is huge.

You know, another concern, in the oil field is always safety. And while it might not be as bad as the, season one of the show, land man might have us believe. I don’t think most people get bags over their head in airplane hangars and are maybe involved in a daily blowout at least. I hope not.

How are ways that, you’re seeing AI lead to better safety solutions, in the patch?

I really think that it comes down to kinda three three areas. Number one, preventative maintenance. So these things are monitoring. You know, AI has the ability to monitor the well conditions, make corrective, you know, tweaks immediately, and and so, you know, prevent these big issues and these big deals. And then it also lets you know what the health of your equipment is. Is it time? And it then it prevents, you know, these massive downtimes that you have, of course, that can cost millions of dollars if something breaks.

You know, there’s a saying about prevention is, but an ounce of prevention is worth a pound on the back end. And so, that’s that’s really helpful. I think the other thing that’s really advantageous is it allows for remote monitoring.

So that way, you you can monitor several wells and several, you know, what’s happening, all at once and not having to drive to each individual location because these wells are spaced as because we’re here in Texas. They’ve gotta be spaced. And so it just it saves from that aspect. I mean, every time something has gone wrong on a well, almost inevitably, I’m like, where was the company man?

He’s always in his truck doing paperwork, and now he can be in his truck doing paperwork and monitoring, you know, what’s happening at these at these different wells. And then, ultimately, it comes down to to really the and and this is the most significance is the human aspect of it and the human safety measures that are there. If it looks like it’s going to blow everybody bail I mean, hopefully, it doesn’t. Right?

But, hopefully, you’re now preventing these catastrophic events both on a human cost and in an environmental cost, and really bringing that down. And safety is ultimately again, it AI is not going to do the work for you. You’re still going to have to have human beings doing that and making it as safe as possible, I think, is the ultimate goal of anybody.

It’ll be interesting to also see if, the railroad commission and other state regulatory agencies are able to leverage some of this in the future for, you know, their well monitoring.

Because as we all know, they’re they’re extremely short staffed. You know, a lot of wells get overlooked. A lot of issues get overlooked for a very long time, can lead to a lot of environmental problems. And so it’ll be interesting to see, you know, how AI solutions get, leveraged in the regulatory space as well.

Yeah. Plus, you know, pros and cons both in the in that situation because it is so detailed and specialized, but I’m sure that, you know, it’s coming and that there will be a way to monitor that.

So, what are some of the, downsides or challenges you see, using AI or with AI really taking off over the next few years?

Yeah. I mean, it really comes back to, you know, how it’s being made and what it’s considering. And if I’m investing a million dollars or more in this, you know, in this well, I wanna know how this is being made. And being told we put it into the computer, it went through the algorithm, and this is the answer that we’re getting. You know, I may depending on your investor or who you’re answering to, they may want a little more specificity to that than other than, I don’t know if this is what the computer told me.

Specificity to that than other than, I don’t know. This is what the computer told me. And so there’s just a believability and a credibility issue, I think, that AI has a little bit right now. Like you were saying, of course, we’ve all heard the horror stories of the hallucinations and, you know, people being sanctioned and all these things. And, you know, it’s just I think it’s a little we’re we’ll get there. It’s just a little bit too soon right now. But I think it certainly can be an advantageous tool relative, to helping make those decisions.

Yeah. And I think we’re gonna see the same thing, on the title examination side, which is, you know, AI might be useful in in sort of high level, data accumulation.

Even, you know, in essence kind of pumping out a maybe a preliminary draft of of a title report or something on attractive land. But I still think that we are, quite some time away from, you know, not needing that human element to come in and really look through the documents, trying to ascertain intent, you know, what the parties were trying to do, looking at the bigger picture.

You know, AI, again, it’s only gonna be as accurate as the source data that’s being pumped into it right now. And until it’s got, you know, a complete universal ability to search, all courthouse records, you know, everything that’s behind a paywall.

You know, I think that you’re not gonna get completely accurate reporting out of that, and it’s gonna continue to be somewhat siloed. You know, maybe maybe a certain broker or a certain courthouse or county is gonna be using maybe an AI solution for their records, but they’re not all necessarily communicating with each other, you know, or at least not yet. You know, we’re not there yet.

So You know, and it brings you up of another really good point about, you know, the security of it all and what happens with the data.

Right? So if I certainly if, you know, I don’t want you to be able to use my data that that I’ve got. Now courthouse records and public records is a different story. That’s available to everybody.

But if I’ve gone out and hired folks and I’ve got an accurate you know, or what I believe to be an accurate beat on something, I don’t want this to be loaded into the AI machine that’s gonna generate answers for you. You you know, you go do your own work. I’ll go do my own work. That’s the nature of, you know, competition and trying to get the trying to get to the wells first.

And so, you know, that’s it’s another struggle and another challenge is, you know, when you when this information goes into the algorithm, then what happens to it? You know? Fine. It gives you the answer, but it keeps it.

And, you know, the way AI works is it continuously takes in information to learn and to build and to grow. Well, I wanna do that in in my vacuum, in my side, and I don’t want it to feed your side. And so how to how to keep those things separate is, you know, another concern and another issue that keeps coming up.

Yeah. And I think certainly protecting those those types of trade secrets and and obviously, for law firms, our our ethical duty of confidentiality, limits in some ways what we can, you know, utilize something like chat GBT for because it is open source, you know, in many ways. And so we can’t just go type in, you know, a lot of confidential client data and ask, you know, ChatGPT or AI to analyze it or summarize it or anything like that. So I can certainly see where, you know, in in analyzing discovery or, you know, summarizing depositions or things like that, it could be helpful. But, again, it’s gonna have to be sort of siloed and and protected and kept confidential in some way, I think, before it it can be very viable.

Molly, I’d like to turn the conversation now to how we specifically see AI being used kind of in the litigation sphere. You know, I’ve heard a lot of of stories about legislatures stepping in, talking about stepping in, things like whether or not AI can be used in jury selection, these types of issues. So what are you kinda seeing from the litigation side?

You know, there’s a lot of, there’s a lot of programs out there that are geared towards litigation, and it’s pretty it’s pretty remarkable in as much as I can load in an original petition, for example, and it will immediately spit out to me, information about what the cause of action is, what the defenses are to this cause of action. It will begin preparing discovery requests, interrogatories, and, requests for production. It will do an analysis of the of the venue, of where is filed, give me intel on the judge, arguably, if it has it.

And all of this is stuff that normally, you know, as first and second year lawyers, I would spend ten, fifteen hours processing, to try to impress my partner that I was working on this. And this will do it in about fifteen seconds. I had a little longer than that. But, so it certainly is expediting those those processes. Again, you’ve gotta have somebody look at it on the back end and say, are these the right questions to be asking? Is this information accurate?

But it is it is leaps and bounds ahead of, you know, prior technologies that we had and really becoming, kind of the mainstream. Now, you know, some of the bigger clients are really looking, are you using this? Because this is a more, effective use of our time than paying associate hours, at ten or fifteen hours versus fifteen minutes, twenty minutes, thirty minutes of this, you know, AI time.

And so we’re seeing it a lot in that way.

You know, speaking about the legislation, there’s a there’s a act pending in the Texas house right now, and I’m going to mess up the name, but it is the Texas act, something about AI. I’ll I can fill in the name for you. We’ll put it in the comments afterwards. But, it’s pending in the house right now.

And it’s really geared at, consumers and consumers interfacing with with, AI and then what businesses are doing on the back end of it. But a lot of it is like you cannot take biometrical information, process it and use it to have kinda outcome predictive measures. Right? And so when I first read it, I was like, man, this would be fantastic if we could do this in picking juries.

If I could take some limited information, put it into this algorithm, and have some sort of information about how you’re going to, you know, ultimately vote as a juror, I mean, god bless. Fantastic. It would make it so much easier. And this bill specifically says you can’t do that.

And so rats will have to stick to the old way a little bit. But, you know, there there is a lot of at least it’s showing a sign that the legislature is recognizing, look, AI is not going anywhere, but how can we keep it from being abusive and how can we keep it from protecting consumers? And right now, it’s all very much geared, towards consumers and not specifically towards, you know, oil and gas or litigation or juries or anything in particular, but showing a very strong sign. There are other bills that are pending in other states.

I think Colorado and Michigan, don’t quote me on these, but other legislatures are starting to put them together.

But I think the other thing is, you know, as as more and more oil companies are using AI, I think it’s an interesting conversation about, does it start redefining the industry standard, and does it start redefining, you know, reasonable, prudent operator? That’s the standard that everybody has held to. And in all the MSAs, it’s you will you will perform this work in compliance with industry standards. Well, does that mean you’re gonna use AI?

Doesn’t mean you’re not gonna use AI. And at some point, the balance is gonna switch. It’s very easy on this side right now to say, no. We’re not we’re not there yet.

And and once we get to the other side, and I believe that we will, it’ll be very easy to say, well, of course, we’re using AI. I mean, this you’ve gotta be insane nuts. It’s like seat belts. Right? You know? It’s like, oh, seat belts were just kinda nice to have. And then now it’s like, of course, we’re using seat belts.

It hasn’t become the industry industry standard yet, but we’re getting there quickly.

And so it’s this interesting transition time of it’s like, okay. If you’re a super major and you’re not using AI, is that should you be? Is that what a reasonably prudent operator does? And then at what point does it trickle down?

And that’s gonna be the standard for, you know, so much of the litigation that follows, certainly, on the negligence side of it, because that’s your the duty as a as a reasonably prudent operator. But even in addressing these contracts of, like, well, you know, would this have taken you this long if you would have used AI? Would this have happened had you been using the AI that’s available to you? And, you know, the the the argument against it is it’s really expensive.

It’s incredibly expensive. And, but then, you know, that’s a slippery slope as well because then are you putting profits over people and all of these narratives that you like to hear from, you know, certainly personal injury attorneys as well. And it’s it’s just really creating this, like, conundrum in this very gray area, which is kind of exciting, to at least there’s an argument to be made about it, from a litigation side of it. But just watching it transition is just it’s really kinda interesting.

Yeah. And presumably, the the price will start to come down. You know, economies of scale will take over. Eventually, more competition.

I mean, we’re already seeing, you know, some of this happening, on Wall Street and, you know, other other countries, coming out with their own answers to chat GPT and and claiming, more efficiency, cheaper, things like that. And I think that’s just the beginning of of where it’s headed. It reminds me a little bit by way of analogy of the effect on the industry that just something as simple as digitizing the courthouse records has had, whereas, you know, you don’t have to send landmen out to all these distant counties for days on end to dig through courthouse records and wait in line, when there’s a boom going on in one area. I mean, I think if any broker’s not using, the online digital records to, pull, you know, oil and gas ownership reports, they’re at a huge disadvantage, you know, and they’re they’re gonna get left by the wayside. And so I think we’re gonna see the same thing, you know, over the next few years with with some of these AI solutions.

Molly, thanks again for joining me today. Thank you everyone out there for joining us for this edition of of OG Talks Good Energy. We’ve really enjoyed putting this one together, and looking forward to seeing what type of dystopian future AI has in store for us.