The difference between pre-bid and post-award protests, and why the former carries a higher success rate. The three-tiered forum system: agency review, GAO, and the Court of Federal Claims, each with different costs, timelines, and levels of scrutiny. How GAO's July 2025 pleading standard change has tightened the path to sustained protests, with only three sustained in the first two months of 2026. The emerging question of whether GenAI is being used in government bid evaluations, and what the FAR's requirement for independent judgment means for contractors. A risk/importance matrix for AI use in proposals and legal filings, grounded in real cases where hallucinated content destroyed contractor credibility.
David also addresses the persistent fear that protesting damages agency relationships, the value of enhanced DOD debriefings, and the critical timelines every proposal team should have mapped before a loss arrives.
David Timm is a partner at Burr & Forman. He chairs the Bid Protest Committee for the Federal Bar Association Government Contract Section and published a widely cited report on AI hallucinations in procurement legal filings in January 2026.
✹ David Timm's LinkedIn: / timmdavid ✹ The 2025 GenAI Misuse Report: https://www.burr.com/government-contr... ✹ Burr & Forman Government Contracts Blog: https://www.burr.com/newsroom/blogs ✹ Implications for the procurement system for NCMA: https://content.ncmahq.org/contractma... ✹ GAO/COFC Splits: https://mcsmag.com/gao-cofc-bid-prote... ✹ Dean Jessica Tillipman (GW Law) on AI in procurement: https://jessicatillipman.com/ ✹ GAO Bid Protest Decisions: http://gao.gov
Christina Carter (00:08) Hey Stargazers, if you have ever lost a federal deal and thought, should we protest this? You already know how tough of a question that is. Most GovCon leaders treat bid protests as a last ditch legal move. Wrong? David Timm lives this every day.
David Timm is a partner at Bernd Foreman, chair of the Bid Protest Committee for the Federal Bar Association Government Contract Section, and he litigates protests and claims before the Government Accountability Office, the US Court of Federal Claims, and the US Court of Appeals for the Federal Circuit. In this episode, David walks us through how to think about a bid protest.
This is all about understanding when and how to protest a decision. Is it an operating risk or is it the right thing to do? Learn how David Timm thinks about bid protests and what they mean for us in GovCon. All right, here's the episode.
Christina Carter (01:03) Hey David, thank you so much for being on the Stargazy Brief. It's wonderful to have you on.
David Timm (01:07) Delighted to be here.
Christina Carter (01:07) For those of you who don't know, David knows about everything there is to know about claims when it comes to bids in US federal contracting.
When I talk with companies, a lot of them say that they think their protests are starting way after they lose. But in your experience, what are the earliest signs? Is it during capture, at Q&A, amendment season?
David Timm (01:28) There are two basic types of protest. One is a challenge to the terms of the solicitation itself. That's called a pre-bid protest. That one tends to have a higher success rate because generally you want to engage with the government through the RFI and Q&A process. If you get an RFP and you're reviewing it and you see something that doesn't make sense to you, it can often have a much higher ROI to clarify that question with the government before you end up bidding.
There are a couple of doctrines depending on whether you go to the Court of Federal Claims or to the Government Accountability Office that prohibit you from protesting the terms of the solicitation after you've bid and the project has been awarded. You need to do that beforehand on the front end. Those tend to be more successful, and the simple answer is because nobody has won yet. There's no intervenor. There's no awardee who has vested interests in making sure they protect the award that they received.
On a post-award protest, the issues range pretty substantially. Generally the GAO's annual report tells us that some sort of inaccurate technical evaluation is the main reason that protests arise. Understanding exactly what you're proposing and how it matches up with what the government is asking for, whether it be past performance or the price you're proposing or the way you're saying you're going to accomplish the work, making sure that is extremely clear in your proposal so there's no room for ambiguity or misinterpretation when the government is doing their evaluation. That can help prevent protests on the front end.
Christina Carter (03:28) Let's say that I'm a growth leader and I'm deciding whether I want to protest a bid or not, or if it's a claims problem. How can a non-lawyer understand the difference between challenging an award or seeking money or time after performance starts, and also recognizing there's no recovery path, there's no point in even trying? How do I figure out my options? Do I have to talk to a lawyer?
David Timm (03:55) I don't think you have to talk to a lawyer, although I'm biased and I think it's good to have a 10 or 20 minute conversation with someone after you lose an award and say, I don't know what I don't know. Is there something here or is this going to be an effort in throwing good money after bad, which we don't want to do.
I would say that eight or nine times out of 10, when someone calls me and they're upset and they say we're thinking of protesting this, I take a look and we generally tell them to save their money for the next proposal. The types of errors that would result in a good result, which is typically either a re-competition or receiving the award of the contract, the paths to get there tend to be pretty narrow because of the way that the law is designed.
A lot of times when people call, and I don't blame them for this, they really wanted to win the award and they thought their proposal was very good and they thought their price was very competitive and they have a lot of ego wrapped up and a lot of vested interests wrapped up in winning. That can have an emotional valence that comes over the decision-making process. It's good to have somebody who has a little bit more objectivity to take a look.
So that's what I think about protests. Now, if you are trying to make some sort of claim for money after the performance of a contract, usually you want to figure out whether you're going to have entitlement very early on. That's probably during performance of the contract. If there's some sort of delay, or if the government is telling you to do work that is out of the scope of your contract, then you need to figure that out quickly and reserve your rights. Otherwise, if you wait too long, you may have forfeited the ability to make that claim on the back end.
I would say that comparing protests and monetary claims, the monetary claims tend to be a little bit clearer. The company knows whether they're making money or not in accordance with the way that they proposed the work and the way that they think the contract is going to go. Disputes arise all the time during actual performance. Sometimes you want to keep the client happy and that makes total sense. You need to pick your battles and figure out, is this going to be a really big deal that's going to impact our profitability? Are we going to make any money at all? Are we going to lose money on the contract?
That's when you need to start seriously asking yourself, have we made the proper reservation of rights? At that point, it might make sense if you've got that inkling in the back of your mind to call somebody, not necessarily me, but somebody that you've got a relationship with that can help you figure those things out.
Christina Carter (06:47) Let's say I lost and I think I should have won. I think I should talk to somebody. I think we do need to protest the bid. What should I be doing in that first 72 hours other than crying to sleep?
David Timm (07:05) It's a stressful thing to have to consider litigation. As a preliminary matter, people tend to be very nervous about protesting because they're worried about it sabotaging their future efforts with the client, with the government agency. I want to say that's a real concern. In some situations, especially if it is an ego-driven protest, that's going to delay the project. A lot of times in post-award scenarios, the government has to stop everything they're doing. The awardee can't move forward with the project until the dispute is resolved. That can cause a lot of hurt feelings, particularly if the government knows and you know that there isn't much to the protest in terms of substance or merit.
But I think generally speaking, the government is big and agencies are big. You need to know your customer. That's the first thing. Are you working with that particular team, with that particular office where the contracting officer is going to be the same for the next 10 projects that you bid? Or is this something that you went after and you weren't planning on working with this particular part of the agency again, and hurt feelings wouldn't be as big of a deal to your future efforts with them?
It's also important to remember that there are many contracting officers even within a specific agency. They turn over all the time.
On top of all that, if you have a legitimate point to make, the government is generally going to respect that you have a legitimate point. Don't forget that the contracting officers and the government want to get things right. A lot of times when you bring a bid protest, you say, hey, we think there was something wrong with the process, and the government says, yeah, I think you're right. Then they redo it. You don't even have to litigate it beyond filing the initial protest.
Going back to your question about what you should do in the first 72 hours, I think this starts with a good internal process for debriefings.
If you've lost the award, you think you should have gotten it. Depending on whether you're doing business with the Department of Defense, you can get an enhanced debriefing. DOD offers enhanced debriefings, which tend to be much better than what is required by the Federal Acquisition Regulations for other agency awards. Each agency, each contracting officer has a different approach in how they communicate with the unfortunate bidders who weren't selected.
That's the first place to start. You need to see what can we do better and what information can we get about why we lost in the first place. After that, it's really important to be mindful of the timelines. If you're going to the Government Accountability Office or the agency itself in an agency protest, the timelines are extremely strict. They tend to be 10 days within the time that you knew or should have known, or within five days of a debriefing. That turnaround can be very difficult sometimes if you don't have your ducks in a row.
I think that involves knowing or having an idea from the debriefing whether there is a there there.
Christina Carter (11:15) How can I know there's something real there for me to protest? Are there top reasons where people should be protesting? If I'm sitting back and I have a process I want to create of saying this is when I protest and this is when I don't, what tends to be some of those red flags I can look at to say we should be protesting, this isn't hurt feelings or ego?
David Timm (11:34) I think it's probably worth spending some time talking with an attorney about your business so that they understand what problems typically arise or could arise when you're in a situation where you may or may not win an award and might have to protest. Depending on what sort of work you do or what sort of contract you're after, the types of protests can vary pretty significantly.
Earlier I mentioned that an improper evaluation of technical tends to be one of the main reasons that protests are sustained. But there are other protests that are easier to resolve and a lot of those tend to deal with whether you satisfied the basic terms of the RFP or the RFQ. If the solicitation required that you have past performance of X, Y, or Z, and you thought that you provided that proper past performance, but they eliminated you or you didn't make it into the competitive range, those discussions tend to be pretty factual. They don't tend to fall into that vague, discretionary place where we see a lot of protests die.
When the Government Accountability Office or the Court of Federal Claims is reviewing protests, they tend to afford a lot of discretion to the agency about decisions related to their contract. After all, they're the person who is going to be the owner and beneficial recipient of the work that you're going to be doing. They're the experts on what they're trying to achieve. GAO and the Court of Federal Claims don't tend to substitute their own judgment for the judgment of the agency.
You want to get out of that gray area. If another competitor is awarded the contract and it's clear that they were not eligible for some reason that you know, because you know your competitors pretty well, let's say it's a set-aside contract for small business and you know that your competitor is not a small business, or this might be a little bit more subtle, the competitor is using a large subcontractor to do too much of the work that violates the limitations on subcontracting or it's a violation of the ostensible subcontractor rule. The fact that you know your competitors pretty well and who they might be working with could be a big advantage in figuring out whether it might make sense to protest.
Christina Carter (14:03) Has anything changed lately, especially in the past year or two of what tends to win or lose in these claims?
David Timm (14:12) That is a great question. GAO recently last year in the middle of the year issued a report and they said that they were clarifying their pleading standard. Before I get into this, I want to back up and talk about the three places that you can protest very quickly. I like to think about it in terms of the major leagues, the minor leagues, and T-ball.
At the top, the place that you're going to get the best review, it's the major leagues, it's the Court of Federal Claims. They're going to be much less deferential to the agency in the decisions that the agency makes compared to GAO or the agency. The minor leagues are going to be GAO. It's a good place to go if you think you can get quick resolution of the protest because it tends to be cheaper. The pleading requirements, the filings that you have to do are much less formal at GAO. On the downside, the timeline for filing is very strict compared to the Court of Federal Claims, which has a more loose standard for timeliness.
That difference in the scrutiny that GAO and the Court of Federal Claims put on the decisions the agency has made is often the difference between winning or losing a protest. That's illustrated nowhere more clearly than when a protester first goes to the Government Accountability Office and loses and then takes a second bite at the apple and goes to the Court of Federal Claims and wins.
I've been in this situation where GAO doesn't give you the full administrative record or they redact too much of the administrative record. Salient facts that you need to make your argument compelling to GAO are not available to you there. But then when you go to the Court of Federal Claims, because the record tends to be much more expansive and the Court of Federal Claims generally wants to see everything, whereas GAO has a 100-day resolution time period that they're trying to work within, a lot of times you can discover things at the Court of Federal Claims you would never have gotten, and it'll result in a win.
The agency is like T-ball and it's very easy for people to understand this intuitively. The agency review is supposed to be at a level above the contracting officer, so it's supposed to be somewhat objective. But at the end of the day, maybe they'll do the right thing. But if they disagree with you at all, I think it's very likely that you're going to lose that protest.
Christina Carter (16:50) Would I always start out at GAO and then kind of move up? Do I have to start out at the agency level and move up? Or can I go straight to the top?
David Timm (17:03) No, you can go to any one of the three. You don't have to start off in the T-ball scenario and then go to the minor leagues and then the major leagues. The reason I mentioned people go from GAO to the Court of Federal Claims is because of the difference in the timeline. Those two forums are not really related to each other. They have independent jurisdiction. You can go straight to the Court of Federal Claims at the outset.
There are a couple of reasons that you might do that. I've written about splits in interpretation between GAO and the Court of Federal Claims where, if you have the exact same facts, GAO will deny your protest because of the way they interpret the Federal Acquisition Regulations and the Court of Federal Claims will say, yes, you win. If you don't know what those splits are, that can be problematic in making that decision. If you go to GAO and lose and then go to the Court of Federal Claims and win, that's great, you ultimately won, but you probably spent much more in legal costs and time because you duplicated the amount of protesting that you did.
Christina Carter (18:22) That was so interesting. I'll get that article from you and put it in the show notes because that's fascinating. I'm going to go back to my original question. Has anything changed in the past couple of years in terms of when you should be doing protests, when you shouldn't be protesting bids?
David Timm (18:44) Thanks for corralling me and bringing me back to your question. I went off on a complete tangent. The reason I was clarifying is because I think it's important to know the differences in the forum. As I think you can tell from the way that I categorize them, I prefer to go to the Court of Federal Claims because you tend to get a more complete review and they're less deferential.
But there are other reasons on top of that. One of those is that GAO recently clarified in July 2025 its pleading standard. If you don't know what that means, put it into plain language: it means that when you go to make your initial filing at GAO, the agency can move to dismiss based on a lack of facts or clear argument in your initial pleading.
If your protest is going to rely on facts that are only in the administrative record, which you only get after a motion to dismiss or a request for dismissal is resolved, you may never get to the administrative record, which means you lose before you even began. The idea there is that you're supposed to cut down on protests that are speculative, that you have an inkling or a vague idea, but you don't have enough facts to show that you deserve to move forward in the process at GAO.
When commentators, including me, were considering whether that would have an impact on protests at GAO and their success rate, there was a split of opinion and a little bit of confusion as to whether they had clarified the pleading standard as opposed to increased the scrutiny. I think it's pretty clear now, we're recording this at the beginning of March 2026, that GAO did not clarify their pleading standard. They tightened it up. That means it's harder to get past an initial agency request for dismissal because GAO wants to see more than mere speculation in the initial filing. That has additional consequences that I think cements the three-tiered difference between GAO, Court of Federal Claims, and the agency, making the Court of Federal Claims continue to be the major leagues.
Christina Carter (21:06) This is why people need to hire you when they want to do this.
David Timm (21:12) The reason it's so clear that GAO has tightened their standard is because there have only been three sustained protests so far in the first two months of 2026. That's lagging pretty significantly behind. Quick stats here: typically, the effectiveness rate of going to GAO is about 50%. The effectiveness rate means either you received some sort of corrective action from the agency or you won the protest on the merits, a sustained decision. There may still be an effectiveness rate of around 50% at GAO so far. We don't know for sure because it's hard to tell.
But because there have only been three sustained protests, it's definitely below the typical 16 to 18% sustain rate that we see there. By comparison, there are many more protests at GAO than at the Court of Federal Claims, and there have been at least three wins at the Court of Federal Claims since the year began.
Christina Carter (22:20) Let's say I do want to protest and I am worried that's going to damage my relationship with the agency. I know a lot of times they switch out who people are. These agencies are large. But have you ever seen that actually hurt the relationship? Or has it not been much of a factor, especially when the protest is for a legitimate reason and not ego-based?
David Timm (22:44) If someone claimed that their relationship with a particular person or a particular part of an agency was ruined due to a protest, I think it would be very hard to prove.
Christina Carter (23:01) I mean not legally. Have you seen them kind of feel that way? Maybe the right question is, do you feel like the worry about the damage to that relationship is usually not founded on anything? Is it more of a worry without actual consequences, or should they be taking care of something when they are protesting to make sure that relationship is sustained?
David Timm (23:32) I think relationships are really important. I don't want to discount the way that people feel if they've had an experience where they think they protested and the agency held it against them in the future. It could be the case. However, the way that the agency would hold it against you in the future is by discriminating against you illegally in a future potential award. That would also be protestable, because the agency is supposed to act fairly in all evaluations, regardless of whether you protested them in a prior evaluation.
I think you're probably right. I want to be careful. I don't want to say that it's completely unfounded to worry about your relationship with the agency, but it's my experience that even folks who protest quite a lot at the same agency, they tend to continue winning projects and they're still in business and I don't see those ill effects.
When I'm involved with a protest, I make sure that there's some merit to it, that there are colorable arguments and that we're putting our best foot forward and that we're not protesting because of ego. Like I said, I tell people probably eight or nine times out of 10, let's save our money and focus on the next award.
Christina Carter (24:49) I hear that a lot, which is why I asked that question. It comes up quite a bit, that we don't want to push back on things like FOIA requests, for example. We don't want to be annoying. But when is it needed? When is it not? Are we being as annoying as we think we are?
So you've argued that debriefings need a new question, and that is: was GenAI used in the evaluation of bids? If the answer is yes, or evasive, what should a contractor do?
David Timm (25:18) This is a really hot topic right now. As you are probably aware, the government is trying through the administration to integrate generative AI, large language models, chatbots into all facets of the government to get some of those efficiency gains that are promised by these products.
While I am currently unaware of a very clear case of use of a tool for evaluation of bids, I think there's no doubt that it's happening in some form or another. On the government side, they're constrained by resources and personnel like contractors are, and I think even more so over the last year with all of the turnover in the government. They're still procuring roughly the same number of dollars, but there are fewer people to do the evaluations. I think it's only fair to believe that a chatbot is being used in some way. Maybe they're summarizing your proposal and then starting from that summary when they're doing their evaluation. That could be a use case. In other cases, it's possible the government is using a specific tool that has GenAI in it to do the full evaluation from soup to nuts.
The FAR requires independent judgment from the government in evaluating a bid. When you're talking about GenAI tools, we know these things are basically black boxes where the very complicated math that creates the outputs can't readily be shown. The work can't be shown. It's like when you do some math by hand and the teacher penalized you for not showing your work. There are no notes that the GenAI tool comes up with, or if you asked it to come up with notes for why it arrived at the conclusion that it did, they wouldn't necessarily correspond perfectly to the output.
I think it's pretty clear that if the government used a GenAI tool or multiple tools to come up with a full evaluation of bids that it would be a protestable issue. I think it's extremely likely, there's no precedent yet, but it's extremely likely that the court or GAO would say, this is not independent judgment. There's a lack of documentation. And as a result, we're going to tell you to reevaluate the bids.
There are official tools that government agencies are entitled to use that are sanctioned by the government. And then there's people using ChatGPT for things all the time, whether or not their office has a policy in place that says they can. I think it's reasonable to expect that folks are using these things maybe only for part of the evaluation, like summarizing a bid.
That wouldn't necessarily be a sustained protest in that case. If we proved they talked to the chatbot for a summary, but then they created their typical notes and they did the evaluation in accordance with the rest of the requirements of the FAR and the solicitation, that would probably be okay. Then there's the full evaluation on the other side and that's not okay. And then there's this whole landscape in between where it isn't clear how a court would decide what level of GenAI involvement is acceptable or unacceptable. It's a very interesting question. There are lots of people who are thinking about it. Everybody is thinking about how to integrate LLMs everywhere.
I highly recommend Dean Jessica Tillipman from GW Law. She's written a lot on this subject. I think she's probably the leading academic on the issue. I would recommend you check out her writing on this.
Christina Carter (29:30) David, are there any similarities between what proposal, capture, and sales teams do on our side and what you do?
David Timm (29:41) There are a lot of similarities. We both deal with deadlines on a constant basis and those are stressful. We also deal with documents where precision is very important. One way I like to think about the use of GenAI in capture that I think is very similar to my practice where I'm making legal filings all the time is this matrix.
You have an X and a Y axis. On one axis there's low risk and high risk. On the other there's low importance and high importance. When things are very important and very risky, for instance a legal filing in a bid protest, then I think the use of large language models in creating that output should either undergo very strict verification processes or not be used at all because of the risk of hallucinations. You can fill in the matrix with your own use cases depending on the sort of work that you do.
You could imagine in the low risk and low importance situation, rewriting a text to a friend to make sure your meaning is conveyed well. The risk of hallucinations or errors there is very low. It's unlikely that facts in a text to your friend about not feeling like going out on Friday are going to be wrong. And even if there were facts that were hallucinated and incorrect, you've got a good relationship with this friend and it won't matter. It's not important.
Then there are other parts of the quadrant where maybe it's very high importance but low risk. I think of this in an initial case analysis situation where I've received a lot of documents from the other side and I want to go through them very quickly. It's a high importance task, but there's very low risk because I'm using the tool to figure out what's here. I'm not immediately taking the output and putting it into a legal filing, which falls into the high risk scenario.
There are a lot of similarities in the way that my profession and yours approaches the use of large language model tools. I think it can be a useful heuristic to think before I use a large language model: am I falling into a risky side of the quadrant? And if so, what verification do I need to make sure I'm not going to get burned?
When I speak about large language models and when I write about them, I try to really narrow the aperture from AI to legal to procurement. In that vein, I've written a report on hallucinations in legal filings from 2025. I covered it all throughout the year. I love it. If you follow me on LinkedIn, I put out posts regularly discussing these issues. I compiled all of my analysis into a report that I published in January. It's really interesting to see how even well-trained lawyers are having issues with accuracy for case citations and facts. Those are making their way into the high risk, high importance part of the matrix.
GAO and the other procurement tribunals are not taking kindly to hallucinations, especially with regard to case law. You may not care when you're submitting an argument whether the case will be used in the future. You only really care whether you win or not. But the way that our legal system works is it's built on precedent, which means that if one decision has a problem with it, and then the next person relies on that same decision, the problem carries through. From a systemic perspective, we care that all of the citations and all of the logic in legal filings is very accurate and factual.
As everyone knows at this point, large language models can hallucinate all sorts of facts. If anybody is telling you that their technology has no hallucinations, they're selling you a bridge to nowhere. I hate to be the bearer of bad news. There are lots of things we can do to try and reduce hallucinations, and the technology gets better all the time.
As the hallucinations get lower, our scrutiny of the outputs gets less. We apply less scrutiny because the outputs look really good. They look very plausible. As the tools get better, the hallucinations don't go away, but they might look even more plausible and there might be fewer of them. There are lots of studies on this, which means we don't look as closely at the outputs. And as a result, we let slip some very important factual errors.
If it's a factual error in the text to your friend about not wanting to go out on Friday, it doesn't matter. But if it's an error in a compliance matrix that you rely on drafting your proposal and you're disqualified from the award without even consideration of your proposal, that's really bad. There are a lot of severe consequences there. You've wasted a lot of money.
You need to understand where you're falling in the matrix. I think it's illustrative to look at how real cases at the GAO and the Court of Federal Claims that involve hallucinations ended up the way that they did.
Christina Carter (35:36) I find that fascinating and I do want people to go and look at the report. I find it fascinating because those hallucinations can be so problematic. I know this is B2B but there is a Forrester report that says they believe legal claims in the US will go up by 20% based on AI hallucinations alone. I have no doubt that that will translate into federal government claims as well.
David Timm (35:58) So far it's hard to disentangle the trend line. At GAO, the number of protests has fallen year over year for a good number of years, maybe 10 years. It's relatively steady, plus or minus 2,000, but it's generally going down. But when I look at all of the cases that I discussed in 2025 and there are five cases so far in 2026 already, through February, when I look at them, most of the filings are by pro se litigants. What that means is that they didn't hire a lawyer and they used ChatGPT or Gemini or Claude as their lawyer.
This is another principle I think is really useful for people to internalize. You have the matrix idea that I put together, but also if you are asking a question from a large language model and you don't know what the answer should look like, if it's in a domain that you're unfamiliar with, like you're not an attorney and you're trying to file a bid protest, and a case comes out and you look at it and you're like, that really supports my argument, that's a very good case. But you don't know if it's a good case or not and you don't take the time to look up the case and you file it, you're going to lose. Every single one of those cases that I've discussed so far that has a hallucination in it ended in a loss for the contractor. Even if the hallucination only gets you warned rather than sanctioned, it undermines your credibility in a way that's very damaging.
There are at least a couple situations where I analyzed the case and I talked with some other attorneys about it and I said, I think if their credibility hadn't been completely shot by the hallucinations, they might have had a chance at winning this.
Christina Carter (37:59) So go hire David if you want to do a bid protest. Don't hire Anthropic to do it.
David Timm (38:06) I also like to say, you don't need to hire me. You can follow some basic principles. If you want to do a pro se protest and you don't want to hire me, I get it. Nobody likes paying lawyers. That's terrible.
But what you want to do is be really careful. If it's going to fall into that high risk, high importance quadrant, then you need to employ special verification procedures. Most of these cases so far have been at the Government Accountability Office. You don't need a Westlaw subscription to look up GAO cases. They're on GAO's website. If you get a case from ChatGPT and it says this is a great case and it supports your argument, all you need to do is read the case. You can go to gao.gov.
That's really all it takes to figure out whether you've got a hallucination. The minimum amount of verification is confirming that the case actually exists.
But then you get into what I was talking about earlier with hallucinations becoming more subtle. They're not restricted to case citations. Sometimes it'll be summarizing part of your proposal and it'll say we included this discussion of X, Y, or Z that you said we didn't include in our proposal and you can find it at pages 14 through 21 of our proposal. Then GAO takes a look at your proposal and it goes to page 13. There are no pages 14 through 21. You never included that information and that's a factual hallucination. You lose. You wasted all that time. You wasted GAO's time. You wasted the government's time.
No one has been debarred or suspended for filing a protest or a monetary claim that has hallucinations in the filings yet. But I've talked to experts in the area. If you're signing your name on something that has factual errors of that magnitude, and you're certifying that all the facts are correct, I think it could potentially be a False Claims Act issue. The consequences might go well beyond the particular dispute that you're in.
Christina Carter (39:57) I'm a huge proponent of AI, but it's not an end-all, be-all silver bullet. So let's say people want to find you. You put out so much work on LinkedIn. You're a wonderful person to follow. Where can people find you? How can we learn more about you?
David Timm (40:18) My law firm, Bernd Foreman, has a government contracts blog where myself and other colleagues publish about timely issues. That's a good place to follow the work that we're focused on.
I think there's a lot of overlap. Your audience is very focused on bids and proposals. I write a lot about bid protests and there's a nexus there. Almost everything that I do, when I speak, when I come on a podcast like this great one, I post about it on LinkedIn. That's probably the easiest way to follow me. You can go to my law firm's website. My phone number's there, my email's there. You can reach out.
Christina Carter (41:12) Thank you so much for being on the Stargazy Brief and we'll have you on again sometime.
David Timm (41:16) Thanks, Christina.
Christina Carter (41:18) Hey Stargazers, thanks for making it all the way to the end. If you found this interesting, if you know somebody who is thinking about doing a bid protest or wants to understand them, please give this a share. Go check out David and his blog posts, his LinkedIn posts. He is so informative, so interesting. You will learn so much on a continual basis and you'll keep up to date on the news and what bid protests are happening right now. All right, I'll see you next week.