Ethical Innovations: Embracing Ethics in Technology

Ethical Innovations: Embracing Ethics in Technology

Menu

Lawyer Sanctioned After AI-Made Fake Cases Exposed

The Alabama Supreme Court dismissed an appeal after finding that the appellant’s filings contained widespread inaccurate legal citations produced through improper use of artificial intelligence. The court said the briefs included numerous fabricated, misquoted, or nonexistent authorities that forced the court and the opposing party to spend substantial time and resources addressing the false or incorrect citations.

The court ordered the attorney who filed the briefs, W. Perry Hall of Mobile, to pay $17,200 in attorneys’ fees and costs, referred him to the Alabama State Bar for possible discipline, and barred him from filing further documents in that court unless an attorney in good standing with the Alabama State Bar signs the filing. A concurring justice cautioned that dismissal is a severe sanction that should be used sparingly and suggested that sanctions for AI-related misconduct generally should be directed at lawyers rather than clients.

The underlying dispute concerns allegations that trustee Bruce Stewart breached fiduciary duties related to two living trusts and involves the appeal of a summary judgment entered in favor of Stewart in Mobile Circuit Court. The court linked the problem of fabricated citations to attorneys’ failure to verify AI-generated legal research.

Original Sources: 1, 2, 3, 4, 5, 6, 7, 8 (plaintiffs) (sanctioned) (costs)

Real Value Analysis

Overall judgment: the article reports a court sanction for a lawyer who used AI-generated fake legal citations, but it gives little practical help to an ordinary reader. It explains what happened and the court’s reaction, yet it does not provide clear, usable steps, deep education about the problem, or practical guidance a typical person can apply right away.

Actionable information The article describes sanctions, a referral to the bar, and a filing restriction, but it does not give clear steps a reader can take. It does not tell lawyers how to verify AI research, provide a checklist for review, direct clients how to respond if their attorney used AI improperly, or show courts’ procedural options beyond this one decision. It mentions the court’s conclusion that attorneys failed to verify AI output, but it stops short of giving concrete methods for verification. For nonlawyers the piece offers no step-by-step actions; for lawyers it only signals a risk without practical mitigation steps. In short, the article lacks immediate, usable instructions or tools.

Educational depth The article conveys the cause and effect at a high level: improper use of AI produced fabricated authorities, which wasted court time and led to sanctions. But it does not explain how AI systems produce hallucinations, what verification procedures would have prevented the problem, or how legal research normally should be done. There are no numbers, statistical context, technical explanation of AI reliability, or broader discussion of policy or precedent that would help a reader understand the systemic risk. The piece is factual but shallow; it reports an outcome without teaching underlying mechanisms or best practices.

Personal relevance The relevance depends heavily on the reader. For Alabama litigants, lawyers, and people worried about AI in legal work, the story is directly relevant because it shows concrete disciplinary consequences. For most ordinary readers the impact is limited: it is an example of professional misconduct but not something that changes everyday choices for most people. The article does not explain how clients can protect themselves or how nonlawyers should adjust behavior, so its practical relevance to the general public is modest.

Public service function The article serves a modest public function by highlighting a legal-system problem and the court’s response. It warns, implicitly, that relying on unverified AI for legal research can lead to harm. However, it fails to offer explicit guidance, safety warnings, or resources for practitioners, clients, or courts. As a result it functions mostly as a news report rather than as public-service guidance that helps people act responsibly.

Practical advice quality There is little to evaluate because the article gives no real “how to.” The one implicit piece of advice is that attorneys should verify AI-generated authorities. But without concrete methods—such as cross-checking citations in primary sources, using established research databases, or documenting verification—the advice is too vague to be realistically followed.

Long-term impact The article can raise awareness that AI hallucinations have real professional consequences, which could affect long-term behavior among lawyers who read it. Yet because it fails to recommend durable changes—standards for AI use, verification protocols, training, or policy changes—it is unlikely to produce sustained improvement on its own. It documents a single severe sanction but does not outline steps that would help others avoid similar problems in the future.

Emotional and psychological impact The reporting is likely to create concern among lawyers and perhaps alarm among clients about the reliability of AI in legal work. Because it provides no constructive next steps, readers may feel frustrated or helpless. It does not offer calming, clarifying advice to reduce fear or demonstrate how to respond effectively.

Clickbait or sensationalism The article is straightforward and not obviously sensationalized. It cites a serious sanction and quotes a concurring justice’s caution about severity. The tone is factual rather than hyperbolic. However, by focusing on the most dramatic outcome (dismissal and monetary sanctions) without practical follow-up, it relies on the newsworthiness of the punishment rather than offering constructive depth.

Missed opportunities The article missed several chances to teach or guide readers. It could have explained why AI hallucinations occur, shown concrete verification methods (e.g., checking citations against official reporters, using database search features, reading cited opinions), offered a simple due-diligence checklist for attorneys and clients, discussed possible ethical or disciplinary frameworks for AI use, or linked to model court rules or bar guidance if available. It also could have suggested how courts might screen filings for machine-generated fabrications or how clients should respond if they suspect misconduct.

Practical, realistic guidance you can use now If you are a lawyer, never rely on AI output without verifying every case, statute, and quotation against an authoritative source such as an official reporter, an authenticated online database, or the primary document itself. Treat AI suggestions as brainstorming, not authority. Create a habit of opening cited opinions or statutes in an official source, confirm exact quotations and page or paragraph citations, and record where you checked them so you can show your verification if questioned. If you must use AI tools for drafting, add a final research step: run each citation through your research database’s citation-check function or manually search the reporter and confirm that the holding and facts cited actually exist.

If you are a client who suspects your lawyer used AI improperly, ask for a written explanation of how the legal research was conducted and request copies or links to primary sources relied on in your filing. If you see questionable or unfamiliar citations, ask your attorney to show the actual cases and where the quotations come from. If you do not get a satisfactory response, consider contacting the state bar or seeking a second opinion from another attorney.

If you are a judge or court staffer thinking about prevention, require filers to certify that all authorities cited were independently verified and consider local rules that require counsel to provide direct links or citations to official reporters for any nonstandard authority. Implement simple screening practices: run suspicious citations through research tools early in the process and give opposing counsel a narrow, early opportunity to flag fabricated authorities.

For nonlawyers trying to assess news like this, compare independent reports from multiple reputable outlets, consider whether the piece explains causes and remedies, and ask what practical protections exist for affected parties. When a report describes professional misconduct, look for follow-up sources that discuss disciplinary outcomes or consumer protections.

These steps are general principles. They rely on basic verification, documentation, transparency, and escalation when answers are unsatisfactory—practical habits that reduce risk even when the underlying article provides no specific tools or protocols.

Bias analysis

"sanctioned the lawyer who filed it after finding that the lawyer’s briefs contained numerous fabricated, inaccurate, and irrelevant legal citations generated by artificial intelligence."

This sentence uses strong words "fabricated" and "generated by artificial intelligence" that push readers to blame the lawyer and AI. It helps paint the lawyer as dishonest and AI as the source of the problem. The wording hides any nuance about why the lawyer relied on AI or whether other factors mattered. It favors a harsh view of the filings without showing the lawyer's intent or context.

"ordered the attorney, W. Perry Hall of Mobile, to pay $17,200 in attorneys’ fees and costs, referred him to the Alabama State Bar for possible discipline, and barred him from filing further documents in that court unless another Alabama State Bar attorney in good standing signs the filing."

This sentence focuses on punishments and names the lawyer, which highlights consequences and makes the account look punitive. It frames the court's actions as decisive and final, which helps the court's authority and may hide any uncertainty about appeals or mitigation. The order of penalties places money first, making financial punishment feel primary.

"The court described the improper use of AI in the plaintiffs’ briefs as widespread and particularly egregious, saying the filings forced the court and the opposing party to spend substantial time addressing fabricated authorities and misquoted or nonexistent cases."

Calling the use "widespread" and "particularly egregious" uses emotional language that magnifies the misconduct. Those words push readers to see the problem as large and severe. They help justify harsh sanctions and do not show evidence or scope, so the claim may overstate how common the issue was.

"A concurring justice noted that dismissal of an appeal is a severe sanction that should be used sparingly and suggested that sanctions for AI-related misconduct generally should target the lawyer rather than the client."

This sentence frames a principle favoring lawyer responsibility over client punishment. It supports the idea that the legal system protects clients and places blame on lawyers. That emphasis helps shield clients and centers professional accountability, which is a choice about who the narrative blames.

"The court emphasized that the problem of fake citations reflected attorneys’ failure to verify AI-produced legal research."

This wording states a cause—attorneys' failure to verify—presented as fact. It assigns blame to attorneys broadly, not only the named lawyer, which can generalize the problem to the profession. The sentence leaves out other causes such as AI design, vendor issues, or workflow pressures.

"The underlying dispute involves allegations about a trustee’s fiduciary duties in litigation over two living trusts, with summary judgment previously entered in favor of trustee Bruce Stewart in the trial court."

Describing the dispute in neutral terms hides details about the parties or stakes that might affect readers' view. The text names the trustee but gives no detail about the plaintiffs, which shifts focus to the trustee and may make the trustee seem like the prevailing, rightful party. This selection frames the dispute minimally and does not show both sides.

"the plaintiffs’ briefs"

Referring to "the plaintiffs’ briefs" repeatedly keeps attention on plaintiff-side filings as the source of error. This phrasing helps make plaintiffs look responsible for the problem and does not equally discuss defense filings, which may shape the reader to see one side as culpable.

Emotion Resonance Analysis

The text conveys a strong tone of disapproval and reproach, visible where the court “dismissed an appeal,” “sanctioned the lawyer,” ordered payment of “$17,200,” referred him for discipline, and “barred him from filing further documents” unless supervised. These phrases express institutional condemnation and convey anger or sternness from the court. The emotional intensity is high because the actions are severe and specific: dismissal, sanctions, financial penalties, referral for discipline, and filing restrictions are concrete, punitive responses. This emotion serves to show the seriousness of the misconduct and to signal the court’s intent to punish and deter similar behavior. Readers are guided to view the lawyer’s actions as wrongful and to feel that corrective measures were necessary.

A sense of alarm and concern appears where the court described the use of AI as “widespread and particularly egregious,” saying filings “forced the court and the opposing party to spend substantial time addressing fabricated authorities and misquoted or nonexistent cases.” Words such as “widespread,” “egregious,” “forced,” “fabricated,” and “nonexistent” carry negative emotional weight and a moderately high intensity. The purpose is to highlight harm and disruption: the misconduct not only violated rules but also burdened others and undermined the legal process. This emotion encourages readers to worry about reliability and integrity in legal work and to support stricter oversight.

A tone of caution and admonition is present where the court “emphasized that the problem of fake citations reflected attorneys’ failure to verify AI-produced legal research.” The word “failure” signals critique and disappointment, at a moderate strength. This frames the issue as preventable human error rather than an inevitable technological problem. The effect is to steer readers toward the belief that professional responsibility and verification are essential, which may inspire practitioners to be more careful and prompt institutions to adopt safeguards.

A tempered voice of restraint and procedural fairness emerges in the concurring justice’s note that “dismissal of an appeal is a severe sanction that should be used sparingly” and that sanctions “generally should target the lawyer rather than the client.” These phrases express concern for proportionality and fairness, with a moderate, measured emotional tone. The justice’s language reduces the harshness of the court’s actions by acknowledging limits and by distinguishing blame. This emotion is used to reassure readers that the judicial system seeks balance and will not punish parties indiscriminately, thereby building trust in the court’s deliberations.

There is an underlying tone of urgency about maintaining legal integrity, particularly in the description of “fabricated, inaccurate, and irrelevant legal citations generated by artificial intelligence.” The triplet of adjectives increases intensity and creates emphasis through repetition, producing a stronger negative emotional response. This rhetorical pattern aims to alarm readers about the multiple dimensions of the problem—fabrication, inaccuracy, irrelevance—and to push for corrective attention. It serves to persuade readers that the matter is serious, multifaceted, and deserving of decisive action.

Finally, a subdued factual detachment is present in neutral details about the underlying dispute—“allegations about a trustee’s fiduciary duties,” “litigation over two living trusts,” and “summary judgment previously entered in favor of trustee Bruce Stewart.” These neutral, specific descriptions carry low emotional intensity and serve to ground the narrative. They help balance the emotive language by providing context and reminding readers that the sanctions relate to a concrete legal dispute. This calming factual tone guides the reader to see the sanctions not as abstract punishment but as part of resolving a particular case.

Overall, the writer uses emotionally charged words—strong verbs like “dismissed,” “sanctioned,” and “forced,” negative adjectives such as “egregious” and “fabricated,” and repetition of related criticisms—to amplify disapproval and concern while also including measured cautions about proportionality. These choices increase the emotional impact, direct reader sympathy toward the court and the opposing party, foster worry about AI-driven errors, and encourage trust in the judicial process through a balancing note about restraint.

Cookie settings
X
This site uses cookies to offer you a better browsing experience.
You can accept them all, or choose the kinds of cookies you are happy to allow.
Privacy settings
Choose which cookies you wish to allow while you browse this website. Please note that some cookies cannot be turned off, because without them the website would not function.
Essential
To prevent spam this site uses Google Recaptcha in its contact forms.

This site may also use cookies for ecommerce and payment systems which are essential for the website to function properly.
Google Services
This site uses cookies from Google to access data such as the pages you visit and your IP address. Google services on this website may include:

- Google Maps
Data Driven
This site may use cookies to record visitor behavior, monitor ad conversions, and create audiences, including from:

- Google Analytics
- Google Ads conversion tracking
- Facebook (Meta Pixel)