AI Bidding Errors: Who Bears the Risk?

Introduction

Artificial intelligence (“AI”) is rapidly making its way into the construction bidding process. Contractors now use AI-powered estimating software to perform quantity takeoffs and analyze costs with unprecedented speed. According to the drafting and engineering software giant Autodesk, estimating teams are increasingly using AI and automation, particularly for quantity takeoffs, cost forecasting, and speeding up bid creation. Yet as digital tools become routine, legal rules governing bids still rely on traditional principles. This raises a pressing question: if an AI tool makes a costly error in a bid, will the legal system treat that mistake any differently than a human error? Courts are only beginning to grapple with AI-related mishaps, but early indications suggest AI errors will be handled much like any other bidding mistake. In other words, contractors will likely be held responsible for errors made by their AI tools, just as they are responsible for the mistakes of human estimators or means and methods under their control.

AI in Estimating and Bidding

AI technology is already being deployed in the construction industry to streamline estimating and bidding. Current tools available in the market use machine learning to automate tasks that were once purely manual. These platforms can scan digital blueprints to measure quantities, compile cost data from past projects, and even optimize bid schedules. For busy estimators, the appeal is obvious: AI can process complex plans in seconds, standardize calculations, and reduce human workload. In practice, AI-assisted estimating promises greater speed and consistency. Contractors can turn around bids faster and potentially bid more projects with the same staff. The productivity gains are real. As noted above, AI tools have been shown to halve the time needed for quantity takeoffs. And by analyzing large datasets (like historical cost databases and real-time material prices), AI may catch trends or cost factors that a human estimator might miss.

However, these advantages come with new risks and uncertainties. AI systems are only as good as their programming and inputs. If the underlying data is flawed or outdated, the AI’s output will be too (“garbage in, garbage out”). There have also been well-documented instances of AI “hallucinations,” where the software confidently produces information that is false or nonsensical. In a bidding context, an AI might misidentify a building element on a plan, omit a scope item, or use incorrect pricing data without any obvious warning to the user. Moreover, many AI models function as a “black box,” meaning they do not provide transparency about how they arrived at a result. An estimator might see the final numbers but have little insight into the AI’s decision process. This opacity makes it hard to trust, yet trust is critical when multi-million-dollar bids are on the line. Even the federal government has acknowledged the growing role of AI in contracting. In a recent memorandum, the U.S. Office of Management and Budget noted that federal contractors “will likely increasingly utilize AI as part of contract performance in situations where the government may not anticipate the use of that AI.” In other words, AI is entering the contracting workflow whether or not the project owner explicitly realizes it.

AI-related bid errors also highlight the tension between digital tools and physical contract documents. While contractors often rely on PDFs or BIM models, most agreements still give printed plans and specifications legal priority. If an estimator bases a bid on outdated or incompatible digital files, the contractor may still be bound by the official hardcopy documents. Standard construction contract forms, such as the AIA and ConsensusDOCS, do not automatically designate BIM models as binding contract documents unless the parties expressly agree. Given these developments, contractors must balance the efficiency gains of AI with its potential pitfalls.

Bidding and Contractor Responsibility

In competitive bidding, a contractor’s bid is generally binding once it is accepted by the owner. Courts have long enforced the rule that contractors are responsible for the means and methods they choose to employ in preparing and performing a contract. The term “means and methods” refers to the tools, techniques, and processes a contractor uses to carry out the work. If a contractor decides to use an AI-driven estimating platform (as one of its methods), the risk of any error by that tool falls on the contractor. In principle, this is no different than if an employee miscalculates a quantity or a subcontractor underestimates a cost: the contractor remains on the hook for the final bid figure. Once the contract is awarded, the contractor must perform the work at the bid price, absent a legally recognized reason to withdraw or reform the bid.

Bidding law does allow relief in limited cases of error, but only for objective clerical mistakes rather than judgment errors. For example, if a bid contained a clear arithmetic error (such as a misplaced decimal point or a transposed digit), courts may permit the contractor to withdraw the bid or reform the contract. Over a century ago, the U.S. Supreme Court recognized that equity can intervene when enforcing a mistaken bid would be unconscionable, as long as the mistake was honest and promptly disclosed. Moffett, Hodgkins & Clarke Co. v. City of Rochester (1900) (allowing a contractor to avoid a bid due to a glaring calculation mistake upon timely notice) is an early landmark case on unilateral bid errors. Since then, many jurisdictions have developed similar doctrines through case law and statutes.

Most public contracting laws provide a procedure for genuine bid mistakes. For instance, California Public Contract Code § 5103 explicitly distinguishes between relievable clerical errors and inexcusable errors in judgment or carelessness. Under § 5103, the contractor must show the error made the bid materially different than intended and occurred in filling out the bid (not a mistake in interpreting the project requirements). Likewise, New York General Municipal Law § 103(11) permits withdrawal of a bid if, within three days of opening, the contractor proves an unintentional substantial arithmetic error or similar clerical mistake that would make enforcing the bid unconscionable. These laws underscore a consistent theme: if the mistake is a simple mechanical slip that any reasonable person could objectively verify (for example, forgetting to carry a digit in an addition, or omitting a line item that is clearly documented in the work papers), relief may be granted. In contrast, errors of business judgment, like underestimating labor productivity or misreading the specifications, are not excusable. Courts have routinely held that a contractor who misjudges the project scope or negligently overlooks requirements must still honor the bid. In short, an honest clerical blunder might be forgiven; a judgment error will not.

Applying these principles to AI, an error generated by an AI estimating tool would likely be analyzed under the same clerical-versus-judgment framework. If the AI’s mistake is truly a clerical or computational glitch, essentially the electronic equivalent of a typo or math error, a court might treat it as a potentially excusable mistake. For example, a Wisconsin court confronted this issue in James Cape & Sons Co. v. Mulcahy (Wis. Ct. App. 2003). In that case, a contractor’s computer software failed to incorporate a last-minute subcontractor price update, resulting in a significant underbid. The court allowed the bid to be withdrawn, finding the error was a mechanical oversight rather than a lapse in professional judgment (especially since the contractor moved quickly to report the mistake). This ruling highlights that even with advanced tools, the age-old distinction persists: purely mechanical errors can merit relief, whereas errors stemming from poor planning or misjudgment do not. Overall, when a contractor opts to use AI in bidding, from a legal perspective, the means-and-methods rule suggests the contractor bears the risk of that choice: any mistake by the AI is effectively the contractor’s mistake, unless it fits the narrow category of excusable blunders.

Image Courtesy of :Michael C Ferri

Who Owns the Error?

When an AI misfires in the bidding process, there is also the immediate practical question of who bears the consequences. From the perspective of the project owner, the answer is straightforward: the contractor is responsible for the bid it submitted. It will not matter to the owner whether the estimate came from an experienced human professional or an algorithm; the bid price is the bid price. If the contractor tries to retract the bid or claims it should not be enforced due to an AI mistake, the contractor faces an uphill battle. As discussed above, relief is only available for a narrow class of excusable errors. An AI error would need to qualify as a clerical mistake that the contractor identified and reported almost immediately. Otherwise, the contractor will likely be held to its bid. In essence, courts are inclined to treat an AI tool as an extension of the contractor’s operations (another “means or method” chosen by the contractor) rather than as an independent actor that can shoulder blame.

Contractors should also be aware that responsibility for bid errors cannot be readily shifted to the AI vendor. While a contractor might try to bring a contract or warranty claim, most software agreements limit liability, and proving negligence in a complex algorithm is difficult. Courts have also rejected treating software developers like licensed professionals such as architects or engineers. In Superior Edge, Inc. v. Monsanto Co. (D. Minn. 2014), the court noted that malpractice claims do not apply to computer consultants. As a result, the legal system does not recognize AI tools, or their creators, as liable parties in bidding disputes. The risk ultimately remains with the contractor who chose to use the tool.

A useful analogy can be drawn from the now notorious Mata v. Avianca, Inc. (S.D.N.Y. 2023), where attorneys were sanctioned for submitting a brief with AI-generated, nonexistent case citations. The court stressed it was the lawyers’ duty, not the AI’s, to ensure accuracy. Similarly, in a 2025 bankruptcy case, a judge flatly stated: “No lawyer should be using ChatGPT or any other generative AI product to perform research without verifying the results. Period.” These episodes illustrate a clear judicial attitude: AI is a tool, and humans are expected to supervise their tools. Applying this to construction, a court is unlikely to accept “the computer made a mistake” as a valid excuse. Contractors will be expected to exercise human oversight over AI-generated estimates, catching any oddities or red flags before bids go out. Failing to do so could be seen as a lack of ordinary care by the contractor. In short, the contractor “owns” the error in the eyes of the law, so it must be proactive in preventing and detecting AI-caused mistakes.

Risk Management and Takeaway

AI can speed up estimating and improve consistency, but it does not change the basic rule: contractors remain accountable for their bids. To manage the risk, contractors should vet AI tools before use, train staff to treat them as aids rather than substitutes, and ensure every AI-generated estimate is reviewed by qualified personnel. They should also confirm how contracts handle discrepancies between digital and printed documents and stay current with emerging guidance from public owners and agencies. Relief for AI-driven errors will be rare and limited to true clerical mistakes; courts expect contractors to supervise technology just as they supervise people. Used wisely, AI can provide a competitive edge, but only when paired with human oversight and sound risk management.

Michael C. Ferri concentrates his practice in the area of transactional construction law. Mr. Ferri represents property owners, developers and lenders in major metropolitan areas. Prior to practicing law Mr. Ferri practiced structural engineering as a professional engineer.

Robert H. Bell focuses his practice on advising highly regulated global companies. He has two decades of experience representing clients on government regulatory and enforcement matters.  Mr. Bell has an active civil litigation practice in state and federal courts. 

Feature Image Courtesy of: Robert H Bell