Law firms hold exactly what attackers want: confidential deal information, litigation strategy, personal and health records, and trust accounts that regularly move large sums. They also run on email, deadlines and trust between people who rarely meet in person. AI has made every one of those weaknesses easier to exploit. Phishing emails now read like they were written by opposing counsel, voices can be cloned from a podcast appearance, and attackers can sort through stolen files faster than ever.
The pressure is visible in incident data. Baker & Hostetler’s 2026 Data Security Incident Response Report, as summarized by the ABA Journal, said the number of incidents its team handled where law firms were the victims nearly doubled in 2025 compared with the more than 30 it handled in 2024. This guide to law firm cybersecurity explains how to protect client data from AI-powered attacks. It covers the threats that hit law firms hardest, including trust account fraud, what the ABA Model Rules require, the controls that matter most and how to choose security vendors.
The Threats Law Firms Face From AI-Powered Attacks
Business Email Compromise Targeting Trust Accounts
The most direct financial threat is wire fraud aimed at client funds. Attackers break into, or spoof, a mailbox belonging to the firm, the client or another party to a transaction. They watch the conversation, then send “updated” wiring instructions shortly before a real estate closing, settlement payout or escrow release. The money goes to a mule account and is moved within hours. The FBI’s Internet Crime Complaint Center reported about $2.8 billion in business email compromise losses in 2024, according to Nacha’s summary of the IC3 annual report, and real estate and legal transactions are a common target.
AI makes these messages harder to spot. Generative tools can match a lawyer’s writing style from past emails, write flawless English in any tone and respond quickly in a live thread. Voice cloning lets attackers follow up with a phone call that sounds like the partner on the matter.
Ransomware And Data Extortion
Ransomware groups target law firms because downtime is painful and confidentiality is everything. Many groups now steal data before encrypting systems, then threaten to publish client files unless they are paid. For a firm, the leak threat can be worse than the outage, since it touches privileged material and client relationships. Our step-by-step guide to handling a ransomware attack covers the first hours of a response.
AI-Generated Phishing And Impersonation
Phishing aimed at lawyers often mimics court notices, e-filing alerts, document-sharing links, opposing counsel or new client inquiries. AI lets attackers personalize these at scale using details from court dockets, firm bios and LinkedIn. Fake client intake is a particular risk: an “overseas client” who needs help collecting a debt, sends a counterfeit check and asks for a refund of the balance from the trust account is an old scam that now arrives with polished documents and responsive email exchanges.
Risks From The Firm’s Own AI Tools
The firm’s own use of AI creates exposure too. Lawyers pasting client documents into consumer chatbots, AI tools connected to document management systems without proper access controls, and vendors that train models on customer data can all lead to unauthorized disclosure. The ABA addressed this directly in Formal Opinion 512, issued on July 29, 2024, which discusses how the duties of competence, confidentiality, communication and reasonable fees apply to generative AI tools.
What The ABA Model Rules Require
Cybersecurity at a law firm is not only good practice. Under the ABA Model Rules of Professional Conduct, which most states have adopted in some form, it is part of a lawyer’s professional duties. Always check your own state’s version, since wording and comments vary.
Rule 1.1, Comment 8: Technology Competence
In 2012, the ABA amended Comment 8 to Rule 1.1 (Competence) to say that, to maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, “including the benefits and risks associated with relevant technology.” Most states have since adopted a version of this duty. In practice, it means lawyers cannot claim ignorance of how email, cloud storage or AI tools expose client information. They need either the knowledge themselves or access to people who have it.
Rule 1.6(c): Reasonable Efforts To Protect Client Information
Rule 1.6(c) states: “A lawyer shall make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client.” The comments to the rule explain that reasonableness depends on factors including the sensitivity of the information, the likelihood of disclosure if additional safeguards are not employed, the cost and difficulty of the safeguards, and the extent to which they would hurt the lawyer’s ability to represent clients.
The standard is reasonable efforts, not perfection. A breach alone does not prove a violation, but a firm that skipped basic, affordable protections such as multi-factor authentication will struggle to show its efforts were reasonable.
Related ABA Opinions
Formal Opinion 477R (2017) addresses securing client communications and supports a risk-based approach to choosing protections. Formal Opinion 483 (2018) covers obligations after a data breach, including acting reasonably and promptly to stop the breach and mitigate damage, and informing clients. Opinion 483 also describes having an incident response plan in advance as a best practice. Formal Opinion 512 (2024) applies these duties to generative AI.
Practical Controls For Law Firms
The controls below give the most protection per dollar for small and midsize firms. Larger firms will add more, but should not skip any of these. For a broader baseline, our small business cybersecurity checklist covers general controls that apply to any office.
Lock Down Trust Account Payments
- Never accept wiring instructions, or changes to them, by email alone. Confirm by phone using a number from the engagement file or another source recorded before the request.
- Tell clients in the engagement letter and at the start of every transaction that the firm will never change wiring instructions by email, and that they should call before sending funds.
- Require two people to approve outgoing wires from trust accounts, with the second approver checking the callback record.
- Use bank controls such as dual authorization in the banking portal, payee allowlists and positive pay for checks.
- Wait for incoming funds to clear fully before disbursing, especially from new clients.
Strengthen Identity And Email Security
- Turn on multi-factor authentication for email, document management, practice management, remote access and banking, preferring phishing-resistant methods such as security keys or passkeys for partners, finance and IT staff.
- Publish SPF, DKIM and DMARC records for the firm’s domains and move DMARC to an enforcement policy.
- Flag external email and warn on look-alike domains that imitate the firm or its clients.
- Alert on new mailbox forwarding rules and sign-ins from unusual locations, both common signs of account takeover.
Prepare For Ransomware
- Keep backups that are offline or immutable, and test restoring them on a schedule.
- Patch internet-facing systems quickly, especially VPNs, remote access tools and file transfer software.
- Run endpoint detection and response on every laptop and server, monitored around the clock by the firm or a provider.
- Limit administrator rights and segment the network so one compromised laptop cannot reach every file.
- Apply retention policies so the firm does not keep closed-matter data longer than needed. Data you no longer hold cannot be stolen.
Govern The Firm’s Use Of AI
- Write an AI use policy that names approved tools and bans entering client information into unapproved ones.
- Review AI vendors’ terms on data retention, model training and access before adoption, consistent with Formal Opinion 512.
- Make sure AI tools connected to document systems respect existing matter access controls and ethical walls.
- Consider whether client consent or disclosure is needed for particular uses, and document the decision.
Train People On The Patterns
Train everyone, including partners, on the patterns behind modern attacks: urgent payment changes, requests to move to a new channel, unexpected document links and voice calls that pressure staff to skip steps. Run phishing simulations that mimic legal workflows, such as fake e-filing notices. Make it clear that pausing to verify a partner’s request is expected, not insubordinate.
Plan The Response Before You Need It
Write an incident response plan that names who leads, which outside counsel, forensic firm and insurer to call, how client notification decisions are made under Rule 1.4 and state breach laws, and how the firm keeps working if systems are down. Run a tabletop exercise at least once a year. Our guide to data breach response plans walks through what the plan should contain.
Choosing Security Vendors
Most firms rely on outside providers for IT and security, and those providers become part of the firm’s reasonable efforts. Rule 5.3 makes lawyers responsible for making reasonable efforts to ensure that nonlawyer assistance, including outside vendors, is compatible with their professional obligations. When choosing a managed IT, security or AI vendor, ask:
- Do you work with other law firms, and do you understand privilege, ethical walls and trust accounting?
- Who can access our data, from where, and how is that access controlled and logged?
- Do you provide 24/7 monitoring and response, or only business-hours support?
- Can you share an independent audit report such as SOC 2 Type II, and when was your last penetration test?
- What are your breach notification commitments to us, in hours?
- For AI vendors: is our data used for training, how long is it retained and can it be deleted on request?
Put the answers into the contract, including confidentiality terms, notification timelines and the right to audit or receive evidence of controls. Check that your cyber insurance covers social engineering and funds transfer fraud, not only system breaches, and that your vendor contracts align with any conditions the insurer sets.
Frequently Asked Questions
Is A Small Firm Held To The Same Standard As A Large One?
The duty is the same, but what counts as reasonable depends on factors such as the sensitivity of the information and the cost of safeguards. A solo practice is not expected to run a security operations center, but it is expected to use affordable basics such as multi-factor authentication, backups and verified wiring procedures.
Can Lawyers Use Generative AI With Client Information?
ABA Formal Opinion 512 does not ban it, but it requires lawyers to understand how a tool handles data and to assess the risk of disclosure before entering client information. Some uses may call for client informed consent. Check your state bar’s guidance, as many states have issued their own opinions.
Who Is Liable If Trust Funds Are Wired To A Fraudster?
It depends on the facts, the jurisdiction and the contracts involved, and disputes often end up in court or with bar regulators. Firms that skipped verification steps are in a weaker position. Clear written warnings to clients and documented callbacks are the best protection.
Do We Have To Tell Clients About A Breach?
ABA Formal Opinion 483 says lawyers have a duty under Rule 1.4 to inform current clients when a data breach affects their information. State breach notification laws and client contracts may add further requirements. Involve breach counsel early.
How Often Should A Law Firm Test Its Security?
At least annually, plus after major changes such as a new practice management system or office move. Testing should include phishing simulations, a review of email and identity settings, and a penetration test of internet-facing systems for firms that handle sensitive matters.
Conclusion
AI has made attacks on law firms more convincing, but the defenses are well understood. Verify every payment instruction through a known channel, require two people to release trust funds, put strong authentication on every account, keep tested offline backups, govern how the firm itself uses AI and plan the breach response in advance. These steps protect clients, satisfy the reasonable efforts standard in Rule 1.6(c) and the technology competence expectation in Rule 1.1, and make the firm a much harder target than the one next door.
