AI Is Making It Cheaper To Sue You and More Expensive To Defend

While all this is still very important, there is another AI risk quietly showing up in courtrooms across the country: AI is making it much easier to sue companies.

The traditional barriers to filing and pursuing a lawsuit are rapidly disappearing. Someone with a grievance, an internet connection, and a generative AI tool can now identify potential causes of action, draft a professional-looking complaint, cite law and statutes, prepare discovery requests, respond to motions to dismiss, and amend pleadings – all without hiring a lawyer.

It is already happening. A 2026 academic study examining approximately 2.8 million federal court filings found that, following the widespread availability of generative AI, the percentage of federal civil cases filed by pro se litigants increased from roughly one in nine to more than one in five. The research also found evidence that AI-assisted pro se complaints increasingly resemble professionally drafted legal work.

There is an important twist, however: the AI-assisted plaintiffs did not appear to win more often. That may sound reassuring. It shouldn't. From a company's perspective, the most important consequence of AI might not be that plaintiffs have become better at winning lawsuits; it's that they are much better at creating and litigating them.

AI Has Changed the Economics of Litigation

The fundamental problem is an economic one. Generative AI has driven the marginal cost of producing legal work toward zero. A plaintiff can ask an AI system:

Draft a federal complaint against my bank arising from this transaction and identify every potentially applicable federal and state law.

Seconds later, the plaintiff may have a 30-page complaint that is ready to file.

Then, after the bank has hired lawyers, investigated its defenses, and filed a motion to dismiss, the plaintiff can ask the AI system:

The bank filed a motion to dismiss. Draft a response addressing each argument.

Another lengthy legal filing appears. The plaintiff can then use the same AI system to drive the litigation forward, for example, asking the tool to:

Draft discovery requests seeking documents relevant to these claims;

Rewrite my complaint to address the deficiencies identified by the defendant; or

Prepare an appeal of the ruling in favor of the defendant.

The plaintiff may spend minutes generating each document. The defendant does not have that luxury. Every filing must be reviewed by counsel. Facts must be investigated. Employees may need to be interviewed. Documents must be preserved and collected. Legal authorities must be researched. Insurance coverage may need to be evaluated. Responses must be drafted. Executives may need to be briefed. A document that costs a plaintiff almost nothing to create can cost a company thousands, or tens of thousands of dollars to address.

That creates a troubling new litigation equation: The cost of generating legal work is approaching zero. The cost of responding to legal work is not.

For companies that interact with thousands or millions of consumers, customers, patients, employees, account holders, or website visitors, that asymmetry matters. A lot.

Yesterday's Pro Se Plaintiff May Look Very Different Today

Companies and their lawyers are being forced to rethink their assumptions about pro se litigation. Historically, many complaints filed without attorneys were relatively easy to identify. They might contain short narratives, obvious pleading deficiencies, limited legal authority, or difficulty articulating recognizable causes of action. Generative AI has changed the appearance of the filing.

A plaintiff can now ask AI to identify the elements of a claim and then draft allegations corresponding to each element. AI can add jurisdictional allegations, anticipate common defenses, cite statutes, organize claims into counts, and make a pleading look remarkably similar to something produced by an experienced attorney.

However, this does not mean the pleading is correct. It may contain factual inconsistencies, irrelevant causes of action, mischaracterized authorities, nonexistent legal theories, or – occasionally – completely fabricated cases. But sophisticated presentation matters. A poorly articulated claim may be relatively easy to dispose of. A professionally structured complaint containing allegations specifically designed to address the elements of multiple causes of action can require substantially more analysis.

The ultimate result may be exactly the same: dismissal. The road to dismissal, however, can become longer and considerably more expensive.

The Biggest Risk Isn't AI Hallucinations

Much of the legal profession's discussion about AI in litigation has focused on hallucinations. Lawyers and litigants have repeatedly been sanctioned after submitting briefs containing nonexistent cases or cases that exist but do not support the law for which they are cited, all generated by AI. Courts have issued standing orders concerning AI use, required certifications, and warned attorneys and pro se litigants that they remain responsible for verifying their submissions. These stories demonstrate real risks and make great headlines. But fabricated cases may not be the most significant long-term risk for businesses.

In fact, the more concerning development may be exactly the opposite: an obviously fake citation can be exposed; a nonexistent case can be identified; and a fabricated quotation can result in sanctions. But AI-generated pleadings are getting better. Imagine a complaint where every cited case actually exists; the statutes are real; the elements of the causes of action are correctly stated; and the complaint contains detailed allegations addressing standing, damages, causation, and jurisdiction.

The underlying claim may still be weak. But determining why it is weak may require significantly more work.

As AI improves, companies should therefore expect the problem to evolve from bad AI-generated litigation toward increasingly sophisticated AI-assisted litigation. That is a much harder problem.

Regulated Companies Have a Bigger Target on Their Backs

This risk may be particularly significant for heavily regulated businesses, including financial institutions.

Consider a customer who has a dispute with their bank. That customer can describe the dispute to an AI system and ask it to identify every potentially applicable law. Within seconds, the customer may receive possible theories involving the Telephone Consumer Protection Act, Electronic Fund Transfer Act, Fair Credit Reporting Act, Truth in Lending Act, Fair Debt Collection Practices Act, privacy laws, unfair and deceptive practices statutes, contract law, negligence, fiduciary duties, and other federal and state theories. Most may not apply. But the customer does not have to know that. The bank does.

The same dynamic exists in health care. AI can transform a relatively routine patient complaint into allegations invoking HIPAA, state medical privacy laws, consumer-protection statutes, breach-of-confidence theories, negligence, contract claims, and other causes of action.

Privacy litigation presents an especially fertile environment. Plaintiffs are using AI to identify website technologies and generate theories under state wiretap statutes, video privacy laws, comprehensive state privacy statutes, consumer-protection laws, and common-law privacy theories.

The risk extends well beyond banking and health care. Insurance companies, employers, retailers and e-commerce businesses, technology companies, telecommunications providers, educational institutions, utilities, transportation companies, automotive businesses, property managers, and hospitality companies all operate in environments where a routine customer, employee, or consumer dispute can implicate multiple federal and state laws. Employment, accessibility, insurance, telecommunications, and other highly regulated industries face similar exposure.

The common denominator is regulatory complexity. Where a single set of facts potentially touches multiple statutes, regulations, or private rights of action, AI can quickly identify and package those theories into a sophisticated-looking complaint. Regulatory complexity used to create a knowledge barrier, but AI is rapidly removing it.

Courts Are Seeing the Same Thing

Judges are beginning to recognize the change. In April 2026, for example, a federal judge in Georgia observed a "marked increase" in the use of generative AI by unrepresented litigants and imposed disclosure and verification requirements on a pro se plaintiff.

Other courts have sanctioned litigants for submitting hallucinated authorities, while judicial systems around the country are considering how existing procedural rules should apply to AI-assisted filings. Those responses are important, but they have limits. Courts can sanction fabricated citations. They can address frivolous claims. They can require litigants to certify their submissions. What courts generally cannot do is prevent someone from using technology to become better at representing himself or herself. And that means businesses should not expect the courts to solve the problem for them.

The Rise of the AI-Powered Serial Plaintiff

There is another potential development companies should be watching closely. Generative AI does not merely make it easier to bring one lawsuit. It makes it easier to bring many.

A motivated individual can reuse prompts, complaints, discovery requests, and briefs across defendants. Once an effective litigation template has been developed, AI can customize that template for different businesses, statutes, jurisdictions, and factual scenarios at minimal incremental cost. That model likely sounds familiar. Companies have already seen variations of it in website accessibility claims, website tracking and privacy litigation, TCPA cases, consumer-protection litigation, and other areas where relatively standardized theories can be asserted against large numbers of defendants.

A sophisticated plaintiffs' firm is no longer necessarily required to build a scalable litigation operation. AI has democratized that model. One motivated individual, with AI, may be able to do it. For businesses, the concern is therefore not simply the AI-assisted pro se plaintiff. It is the potential emergence of the AI-powered serial plaintiff.

Winning Can Still Be Expensive

Suppose a customer brings a claim involving a $500 transaction. The company believes the claim has no merit. The plaintiff uses AI to produce a 45-page complaint asserting eight causes of action. The company hires counsel, who investigates the underlying transaction, interviews employees, analyzes eight causes of action, researches the authorities cited in the complaint, evaluates jurisdiction and standing, and prepares a motion to dismiss.

The plaintiff then uses AI to generate a 35-page response. The court dismisses six claims but permits two to proceed. The plaintiff sends 40 document requests, 25 interrogatories, and requests for admission. Months later, the company wins. The company was right. The plaintiff recovers nothing. And the company may still have spent tens of thousands of dollars proving it.

That is the business problem AI creates. A meritless lawsuit does not have to succeed to impose substantial costs.

What Companies Should Be Doing Now

Companies cannot prevent customers, patients, employees, or other individuals from using AI. But they can, with the guidance of experienced outside counsel, adjust their litigation-risk strategies to account for the new reality.

1. Identify the disputes most susceptible to AI escalation.

Companies should examine recurring complaints that historically involved relatively low litigation risk: billing disputes, account closures, fraud claims, privacy complaints, accessibility issues, employment disputes, consumer requests, reimbursement disputes, and similar matters.

Ask a new question: If someone ran this dispute through an AI system and asked, "What can I sue this company for?" what would the answer look like?

That exercise may reveal risks traditional complaint-management processes overlook. Experienced outside counsel can help companies identify the legal theories AI is most likely to generate and assess which disputes warrant early legal involvement.

2. Revisit the economics of early resolution.

Some companies have relatively rigid approaches to low-dollar demands because historically there was little reason to believe the claimant would actually litigate. That assumption may need to be reconsidered. The relevant calculation is no longer simply, "What is this claim worth?" but rather, "What will this claim cost if AI enables this individual to litigate it aggressively for the next 12 months?"

That does not mean companies should pay meritless claims. Doing so can create its own incentives for serial litigation. It does mean the economics of early resolution should be evaluated with AI-enabled litigation costs in mind. Outside counsel who understand both the substantive law and the cost dynamics of modern litigation can help calibrate those decisions.

3. Preserve the evidence that makes weak cases easy to defeat.

When plaintiffs can generate allegations quickly, defendants need to be able to establish the facts quickly. Transaction records, customer communications, consent records, call recordings, system logs, privacy records, policies, complaint histories, and other contemporaneous documentation become increasingly important.

The best defense to a 40-page AI-generated narrative may be three pages of indisputable business records.

4. Develop an AI-assisted plaintiff playbook.

Legal departments should consider establishing procedures for identifying and managing unusually prolific pro se litigants. The playbook might address escalation criteria, litigation holds, centralized communications, repeat-plaintiff tracking, early dispositive strategies, settlement authority, insurer notification, regulatory implications, and circumstances warranting consideration of sanctions or vexatious-litigant remedies. Experienced outside counsel can help develop these playbooks and provide the institutional knowledge to apply them consistently across matters.

Consistency becomes particularly important when the plaintiff can generate communications faster than the organization can respond to them.

5. Look for patterns, not just individual lawsuits.

Companies should track pro se demands and litigation across the organization:

• Are multiple plaintiffs asserting the same unusual legal theory?
• Are complaints using identical language?
• Are certain statutes suddenly appearing in customer disputes?
• Are particular jurisdictions seeing increased filings?
• Is the same plaintiff suing multiple companies?

The ability to recognize patterns early can materially change defense strategy. A knowledgeable outside counsel who tracks emerging litigation trends and coordinates defenses across clients can be an invaluable resource in identifying these patterns before they become costly.

6. Know when to push back.

Pro se status is not a license to abuse the judicial process. Depending on the facts and jurisdiction, businesses defending AI-generated lawsuits should consider Rule 11 and state equivalents, statutory fee-shifting, motions to strike, protective orders, limitations on abusive discovery, vexatious-litigant procedures, prefiling injunctions, and other available remedies.

Those tools should be used strategically. But companies should not assume that every AI-assisted pro se litigant must simply be accommodated indefinitely. Outside counsel with experience in sanctions practice and vexatious-litigant procedures can help companies determine when these remedies are appropriate and pursue them effectively.

A New Litigation Risk That Isn't Going Away

The first wave of discussion about AI and litigation focused on lawyers using ChatGPT and accidentally citing imaginary cases. That was probably just the opening act.

The bigger story may be what happens when millions of people gain access to a tool capable of explaining legal rights, identifying potential claims, drafting pleadings, generating discovery, and responding to legal arguments around the clock.

AI does not turn every grievance into a viable lawsuit. It does not make bad facts good. And it does not transform a meritless legal theory into a winning one. But it does something potentially just as consequential: It dramatically reduces the friction required to turn grievance into litigation.

For businesses, particularly financial institutions, health care organizations, and other companies operating in heavily regulated environments, that changes the risk calculation. The next litigation threat may not come from a new plaintiffs' firm, a new statute, or a new theory of liability. It may come from an individual sitting at home who discovered that, for the first time, generating a sophisticated-looking lawsuit costs almost nothing.

AI is making it cheaper to sue. Companies need to take precautions now, including building relationships with experienced outside counsel who understand these emerging risks, to make sure it does not make them unnecessarily expensive to defend.

If you need help assessing these risks or developing processes and procedures to protect your company from AI-assisted pro se litigants, please contact Matthew G. White, AIGP, CIPP/US, CIPP/E, CIPT, CIPM, PCIP or Robert F. Tom, or any member of the Artificial Intelligence team.