AI Errors and Professional Liability: What New York Professional Services Firms Need to Review
New York professional services firms are adopting ChatGPT, Microsoft Copilot, Google Gemini, Claude, AI-enabled CRM systems, document-review platforms, accounting applications, and industry-specific tools. These systems can improve speed and productivity. They can also produce incorrect recommendations, fabricated citations, incomplete analysis, biased results, confidentiality breaches, and content that creates intellectual property disputes.
The central insurance question is direct:
If a client relies on an AI-assisted recommendation and suffers a financial loss, which insurance policy responds?
The answer depends on what went wrong, who suffered the loss, how the service was delivered, and the exact wording of your policies. Professional liability or errors and omissions insurance may address a claim arising from negligent professional services. Cyber insurance may address a privacy or network-security incident. Media or intellectual property coverage may apply to certain content claims. Crime insurance may respond to theft or social-engineering fraud. General liability usually does not replace professional liability for an incorrect recommendation.
This article explains the coverage issues New York lawyers, accountants, consultants, brokers, architects, engineers, technology firms, and other professional services businesses should review.
If you are reading this as part of our generative AI insurance series, start with Generative AI, Insurance, and Business Risk: A Practical 2026 Guide for Connecticut and the Northeast. It provides the broader framework for evaluating AI-related errors, privacy risks, intellectual property exposure, and insurance gaps.
Important answer: New York professional services firms should not assume that an AI-related claim is automatically covered or excluded. Review professional liability, cyber, media/IP, crime, general liability, business interruption, and contract provisions together. The policy language: not the name of the technology: controls the analysis.
Why AI creates a professional liability problem
Professional services firms have always been responsible for the accuracy, suitability, and quality of the services they provide. Generative AI does not change that responsibility.
AI changes the way an error may occur. A professional may use a tool to:
Summarize a lengthy contract or financial report
Identify potential legal authorities
Draft client correspondence
Recommend an accounting treatment
Analyze an engineering or architectural document
Create a risk assessment
Screen or prioritize client information
Produce marketing copy, images, or presentations
Answer client questions through a chatbot
Compare insurance policies or identify exclusions
Automate a CRM workflow
Generate a first draft of a report, opinion, or proposal
The tool may produce an answer that appears confident and complete. That appearance can create a dangerous gap between the professional’s actual review and the client’s reasonable reliance.
An AI system may:
The insurance claim often begins with the professional service. It may then expand into a privacy, cyber, intellectual property, regulatory, or contractual dispute.
What New York DFS Circular Letter No. 7 means: and what it does not mean
New York Department of Financial Services Circular Letter No. 7 (2024) addresses the use of artificial intelligence systems and external consumer data and information sources by insurers.
The guidance applies to insurers authorized to write insurance in New York and focuses on the use of AI and external data in insurance underwriting and pricing. It addresses subjects such as:
Governance and senior-management oversight
Model validation and monitoring
Fairness and non-discrimination
Actuarial support
Proxy discrimination
Consumer transparency
Third-party vendor management
Documentation and recordkeeping
Circular Letter No. 7 is important for the insurance marketplace. It is not an automatic coverage rule for a New York law firm, accounting practice, consulting firm, brokerage, architecture firm, engineering firm, or other business using AI.
It does not automatically:
Add AI coverage to your professional liability policy
Exclude AI-related claims from your policy
Require every business to purchase a separate AI liability policy
Decide whether an AI error is professional negligence
Replace the definitions, exclusions, conditions, and endorsements in your policy
Transfer responsibility from your firm to the AI vendor
The distinction matters. The circular regulates how insurers use AI and external consumer data in underwriting and pricing. Your firm’s coverage depends on the policy contract issued to your business.
You should still ask whether your insurer is using AI or external data in underwriting your account and whether the insurer has added AI-related questions, exclusions, warranties, or endorsements at renewal. However, do not treat Circular Letter No. 7 as a guarantee that your firm’s AI-related claim will be covered.
Professional liability and E&O: the primary policy for negligent advice
Professional liability insurance, also called errors and omissions or E&O insurance, is designed for claims alleging that professional services caused a client financial harm.
Typical allegations may include:
Negligent advice
An incorrect recommendation
Failure to identify a material issue
Missed deadlines
Inaccurate analysis
Failure to perform promised services
Failure to meet the applicable professional standard
Misrepresentation connected to professional services
A client’s financial loss after relying on the firm’s work
For AI-related claims, E&O coverage may be relevant when the client alleges that the firm failed to review, verify, supervise, or correct an AI-assisted work product before delivering it.
Example: New York law firm
A New York law firm uses an AI tool to identify relevant case law. The system generates a fabricated citation. An attorney includes the citation in a filing without independently checking it. The court rejects the filing, the client loses a procedural opportunity, and the client alleges that the firm’s failure caused financial harm.
The issue is not that the lawyer used AI. The issue is whether the lawyer or firm delivered negligent professional services. A professional liability policy may be relevant, subject to the policy’s terms, exclusions, retention, limits, and claims-made reporting requirements.
Example: accounting firm
An accounting firm uses an AI-enabled tax analysis platform to prepare a client recommendation. The tool incorrectly interprets a transaction and fails to account for a relevant limitation. The client follows the recommendation, incurs additional tax liability, and alleges professional negligence.
This is the type of fact pattern that may fall within the conceptual territory of E&O insurance. Coverage still requires a policy-specific review. Some policies may contain exclusions affecting tax advice, technology services, contractual liability, or claims based on guarantees.
Example: insurance broker
An insurance broker uses an internal AI assistant to summarize a commercial policy. The assistant fails to identify an exclusion for a particular business activity. The broker tells the client that coverage exists. After a loss, the carrier denies the claim, and the client seeks damages from the broker.
The client’s allegation focuses on professional advice and a failure to identify a coverage limitation. Professional liability is central. The broker should also review cyber coverage if confidential policyholder data entered the system and crime coverage if the incident involved fraudulent instructions or funds transfer.
What to review in your E&O policy
Ask your broker and coverage counsel to review:
The definition of “professional services”
Whether AI-assisted work falls within that definition
Technology services exclusions
Software and system-performance exclusions
Contractual liability exclusions
Bodily injury and property damage exclusions
Prior knowledge provisions
Fraud and intentional misconduct exclusions
Intellectual property exclusions
Privacy and confidentiality provisions
Defense-cost treatment
Consent-to-settle language
Claims-made and reported requirements
Retroactive dates
Extended reporting options
Sub limits and special endorsements
Any AI-specific exclusion, limitation, or warranty
Many professional liability policies are written on a claims-made basis. That means the timing of the claim, the policy period, the retroactive date, and the reporting process can affect coverage. Do not wait until a renewal or lawsuit to determine how a potential AI incident must be reported.

Cyber insurance: important for confidentiality and system incidents
Cyber insurance generally addresses risks arising from data breaches, network security failures, cyberattacks, ransomware, privacy events, and related business interruption. Coverage varies significantly between insurers.
Cyber coverage may address:
Incident response
Breach counsel
Forensic investigation
Notification expenses
Credit monitoring
Data restoration
Cyber extortion
Network interruption
Certain regulatory investigations
Privacy liability
Security liability
Some forms of social engineering or funds-transfer fraud
Cyber insurance becomes relevant when an AI event involves the unauthorized disclosure, loss, theft, or misuse of information.
Example: confidential client information enters an AI platform
An employee copies a client’s confidential financial records into a public AI tool to generate a summary. The firm later learns that the information was processed or retained in a manner inconsistent with the client agreement.
The client may allege:
Breach of confidentiality
Negligence
Contract violation
Privacy-law violations
Failure to safeguard information
Professional negligence
The professional liability and cyber policies may both matter. The cyber policy may address the privacy or security event. The E&O policy may address allegations that the firm mishandled information while delivering professional services. The policies may contain overlapping-insurance provisions, exclusions, or allocation language.
Cyber insurance does not automatically cover bad advice
A cyber policy is not a substitute for E&O. If an AI tool produces incorrect legal, accounting, engineering, tax, insurance, or consulting advice, and the client suffers a financial loss without a data breach or network-security event, cyber coverage may not respond.
Review whether your cyber policy contains:
Privacy liability coverage
Network security liability coverage
Coverage for vendor-related incidents
Coverage for wrongful collection or use of data
Social engineering coverage
Broad definitions of “computer system” and “network”
Business interruption coverage triggered by a system failure
Exclusions for professional services
Exclusions for contractual liability
Exclusions for voluntary disclosure of information
Requirements for MFA, backups, endpoint protection, and employee training
Your firm’s AI governance can affect underwriting. Insurers may ask how you control access to AI tools, whether employees receive training, whether vendors retain prompts, and whether your firm uses sensitive information in external platforms.
Media and intellectual property insurance: review AI-generated content
Professional services firms increasingly use AI to create:
Website content
Advertising copy
Social media posts
Client presentations
Illustrations
Marketing images
Video scripts
Training materials
Software code
Reports and publications
An AI-generated or AI-assisted work product can create intellectual property exposure. A third party may allege copyright infringement, trademark infringement, misappropriation, right-of-publicity violations, or unauthorized use of protected content.
Media liability or intellectual property coverage may be available through:
A media liability policy
An advertising injury grant
A cyber policy endorsement
A technology E&O policy
A professional liability policy endorsement
A stand-alone intellectual property policy
Do not assume that general liability’s “personal and advertising injury” coverage will address every AI-generated content claim. Definitions and exclusions vary. Some policies may exclude claims involving professional services, contractual promises, intentional acts, known infringement, or technology-related activity.
Practical content controls
Before using AI-generated content for a client or your own firm:
Identify the source and purpose of the output.
Require human review by someone who understands the subject.
Check quotations, citations, statistics, names, and legal authorities.
Conduct trademark and copyright screening where appropriate.
Keep records showing who reviewed and approved the content.
Confirm that the client engagement permits AI-assisted work.
Review vendor terms regarding ownership, indemnification, training use, and output rights.
Avoid representing that content is original unless you have a reasonable basis to make that statement.
Insurance supports risk management. It does not replace clearance procedures or professional judgment.
Crime insurance: when the problem involves theft or fraud
Crime insurance addresses dishonest acts and certain forms of financial crime. Depending on the policy, it may cover employee theft, forgery, alteration, computer fraud, funds-transfer fraud, or social engineering.
Crime coverage may become relevant when:
An AI-enabled workflow changes payment instructions
A criminal uses a deepfake voice to impersonate an executive
An employee uses an AI tool to facilitate a fraudulent transfer
A compromised account sends unauthorized payment instructions
A vendor’s system is manipulated to redirect funds
Crime insurance generally does not cover a client’s financial loss caused only by an incorrect professional recommendation. It also may not cover every social-engineering loss unless the policy includes the right insuring agreement and the firm follows required verification procedures.
Review:
Social-engineering sublimits
Callback and verification requirements
Dual-approval controls
Coverage for fraudulent instructions
Coverage for vendor impersonation
Computer fraud wording
Exclusions for voluntary parting with property
Employee dishonesty provisions
Definitions of “financial institution” and “computer system”
Your firm should use independent verification procedures for payment changes, regardless of whether AI is involved.
General liability: useful, but usually not the answer to an AI advice claim
General liability insurance typically addresses third-party claims involving bodily injury, property damage, and certain personal or advertising injuries.
For a professional services firm, general liability may respond to an unrelated incident such as:
A visitor slips in your Manhattan office
An employee accidentally damages a client’s physical property
A contractor causes property damage at a client location
A visitor alleges certain covered advertising injury
General liability usually does not function as professional liability insurance. A client alleging that your AI-assisted tax recommendation, architectural analysis, legal memo, consulting report, or insurance advice caused financial harm will generally require an analysis under professional liability coverage rather than relying on general liability.
Review the professional services exclusion carefully. It may limit or exclude losses arising from the performance of professional services.
Business interruption: the trigger matters
AI-related disruption can cause a firm to lose revenue. A platform outage, corrupted data, ransomware event, vendor shutdown, or unusable software environment may interrupt operations.
Business interruption coverage may exist under:
A commercial property policy
A business owners policy
A cyber policy
A technology policy
A contingent business interruption endorsement
The key question is what triggered the interruption.
A cyber policy may respond to an interruption caused by a covered network security event. A property policy may require direct physical loss or damage to covered property. A BOP may contain business income coverage tied to a covered property loss. None of these policies necessarily cover lost income because an AI tool generated a poor recommendation, delayed a project, or required internal rework.
Ask about:
The covered trigger
Waiting periods
Period of restoration
Dependent business interruption
Service-provider interruption
System failure coverage
Data restoration
Extra expense
Contingent technology failure
Sublimits
Contractual service-level remedies
A firm that depends on an AI vendor for document processing, client intake, analysis, or communications should understand whether a vendor outage qualifies as a covered event.

Contractual risk transfer: useful protection with limits
Contracts can allocate certain AI-related responsibilities between your firm, your client, and your vendors. Contractual risk transfer may include:
Indemnification provisions
Defense and hold-harmless obligations
Limitation-of-liability clauses
Client approval requirements
Disclaimers regarding AI-assisted work
Confidentiality provisions
Data-processing terms
Security requirements
Vendor indemnification
Intellectual property warranties
Insurance requirements
Additional insured requirements where appropriate
A contract does not automatically create insurance coverage. Many policies exclude liability assumed solely under contract, although coverage may exist for liability the insured would have had without the contract.
Your firm should avoid copying generic AI language into every engagement letter. The clause must reflect the actual service and risk.
For example, a consultant providing an AI-assisted market analysis should not promise that every recommendation is error-free. A law firm should not disclaim the professional duty to review work product. An architect or engineer should not use a disclaimer to avoid obligations imposed by professional standards, building requirements, or the engagement itself.
Vendor contracts also deserve close attention. Review whether the AI provider:
Uses prompts or uploaded documents to train models
Retains data
Provides security commitments
Offers an indemnity for intellectual property claims
Limits liability to a small subscription fee
Disclaims accuracy
Disclaims consequential damages
Permits subcontractors
Provides audit or incident-notification rights
Offers an enterprise privacy configuration
The vendor’s indemnity may be narrower than your client’s claim against your firm. A vendor contract cannot replace your own E&O and cyber program.

Coverage comparison for AI-related claims
Risk or allegation | Most relevant coverage to review | Why |
Incorrect AI-assisted legal, tax, accounting, engineering, or consulting advice | Professional liability/E&O | Addresses alleged errors, omissions, negligence, or financial harm arising from professional services |
Confidential client data entered into an AI platform | Cyber and professional liability | May involve privacy, security, confidentiality, and professional-service allegations |
AI-generated marketing content allegedly infringes copyright | Media/IP, cyber, or advertising injury coverage | Coverage depends on the wording, content type, and applicable exclusions |
AI tool produces a biased recommendation | E&O, cyber, regulatory, and management liability | The claim may involve negligent professional services, privacy, discrimination, or regulatory allegations |
AI-enabled fraudulent payment instruction | Crime and cyber | Crime may address funds-transfer fraud or social engineering; cyber may address the compromise |
Office visitor injured during an AI-related meeting | General liability | Bodily injury is generally a general liability issue, subject to policy terms |
AI platform outage stops client work | Cyber, business interruption, or technology coverage | The policy must contain the correct system-failure or interruption trigger |
Client alleges breach of an AI-related contract promise | E&O and contract review | Coverage depends on whether the claim involves covered negligence or only assumed contractual liability |
AI-generated software causes client loss | Technology E&O, professional liability, and cyber | The definition of professional services and technology exclusions become critical |
This table is a starting point, not a coverage determination. One claim can involve several allegations and several policies.
A New York AI insurance review checklist
Use this checklist before your next renewal or whenever your firm expands AI use.
1. Inventory every AI tool
List:
ChatGPT
Copilot
Gemini
Claude
AI CRM features
Document-review systems
Accounting and tax platforms
Design and engineering applications
Client-facing chatbots
Automated intake systems
Internal knowledge tools
Vendor systems that process your firm’s data
Record the purpose, users, data access, vendor, retention settings, and whether the tool produces advice or only administrative assistance.
2. Classify the output
Determine whether the tool creates:
Internal administrative material
Client-facing communications
Professional recommendations
Legal or regulatory analysis
Financial or tax advice
Engineering or architectural work product
Marketing or advertising content
Code or technical deliverables
Decisions affecting individuals
The greater the client reliance and professional consequence, the more important your E&O review becomes.
3. Establish human approval
Require qualified human review before AI-assisted work reaches:
A client
A court
A regulator
A lender
An investor
A project owner
A policyholder
A third-party recipient
A disclaimer stating that AI may be inaccurate does not replace review.
4. Protect confidential information
Define what employees may and may not enter into AI tools. Your policy should address:
Client names
Social Security numbers
Financial records
Health information
Litigation materials
Trade secrets
Credentials
Policyholder information
Proprietary designs
Unreleased transactions
Document approved tools and prohibit unapproved public platforms for sensitive information.
5. Review your insurance program as one system
Ask your broker to coordinate:
E&O or professional liability
Cyber liability
Media and intellectual property coverage
Crime insurance
General liability
Business interruption
Technology E&O
Employment practices liability, if AI affects hiring or employment decisions
Directors and officers liability, if governance failures create management claims
A policy-by-policy review can miss overlaps and exclusions.
6. Preserve evidence
Keep records of:
The prompt or instruction
The source documents used
The AI output
Human edits
Reviewer identity
Approval date
Final version
Client disclosures
Vendor terms
Incident reports
These records help your firm investigate an error and respond to an insurer’s questions.
What New York professional services firms should ask their broker
Ask these questions in writing:
Does the professional liability policy cover claims arising from AI-assisted professional services?
Is there an AI exclusion, limitation, warranty, or sublimit?
Does the definition of professional services include technology-enabled advice?
Does the cyber policy cover unauthorized disclosure to an AI vendor?
Does cyber coverage apply when an employee voluntarily uploads confidential information?
Is media/IP coverage available for AI-generated text, images, code, or designs?
Does crime insurance cover AI-enabled social engineering?
What business interruption trigger applies to an AI vendor outage?
Are vendor errors excluded?
What notice must the firm provide after discovering a possible error?
Is the policy claims-made and reported?
Are defense costs inside or outside the liability limit?
Are contractual indemnity obligations covered?
Do client contracts require specific limits or endorsements?
What controls will the underwriter expect at renewal?
Bring your actual policy forms, endorsements, client contracts, vendor agreements, and AI-use policy to the review. A general verbal statement that “AI is covered” is not enough.
A note for lawyers, accountants, brokers, architects, engineers, and consultants
The risk differs by profession.
Law firms should verify citations, authorities, quotations, deadlines, privilege handling, and client confidentiality.
Accounting firms should validate calculations, tax positions, reporting requirements, source data, and professional judgments.
Insurance brokers should confirm coverage grants, exclusions, endorsements, limits, conditions, and policy comparisons independently.
Consultants should validate assumptions, forecasts, recommendations, and deliverables against the engagement scope.
Architects and engineers should review specifications, calculations, drawings, code references, safety considerations, and project requirements.
Technology firms should evaluate software performance, security, output accuracy, IP rights, and client dependency.
Marketing and communications agencies should screen AI-generated content for copyright, trademark, privacy, publicity, and advertising claims.
The professional remains accountable for the final service delivered to the client.
Frequently asked questions
Does New York DFS Circular Letter No. 7 require my firm to buy AI insurance?
No. Circular Letter No. 7 provides guidance to insurers regarding their use of artificial intelligence systems and external consumer data in underwriting and pricing. It does not automatically require New York professional services firms to buy AI insurance or establish coverage under a business policy.
Does professional liability insurance cover AI errors?
It may, if the claim alleges covered negligence, an error, an omission, or another covered failure in professional services. Coverage depends on the policy wording, exclusions, claims-made conditions, and facts of the incident.
Is cyber insurance enough if my firm uses ChatGPT or Copilot?
No. Cyber insurance is primarily designed for cyber incidents, privacy events, network security failures, and related losses. It does not automatically cover an incorrect legal, accounting, engineering, consulting, or insurance recommendation.
What happens if an AI tool misses a policy exclusion?
If a client relies on the professional’s incorrect coverage advice and suffers a financial loss, the allegation may involve professional negligence. A broker should review E&O, cyber, crime, contract, and policy-specific exclusions.
Can a disclaimer eliminate liability for AI mistakes?
No. A disclaimer may help define expectations, but it does not automatically eliminate professional duties, contractual obligations, statutory requirements, or insurance claims.
Should New York law firms prohibit generative AI?
Not necessarily. A blanket prohibition may be unnecessary, but firms need documented governance, approved tools, confidentiality controls, human review, training, and clear responsibility for final work product.
Does general liability cover a client’s financial loss from bad AI advice?
Usually, general liability is not the primary policy for that exposure. Professional liability or E&O is generally more relevant to claims arising from negligent professional services.
Can an AI vendor’s indemnity replace professional liability insurance?
No. Vendor indemnities may be narrow, capped, conditional, or subject to exclusions. Your firm remains responsible for its client relationship and should maintain its own insurance.
How often should a New York firm review AI-related insurance?
Review coverage before implementing a new client-facing tool, uploading sensitive data, expanding AI into professional recommendations, signing a major client contract, or renewing your insurance. Conduct a formal review at least annually.
Review your New York professional liability program before a claim
AI adoption is not only a technology decision. It is a professional-services, contract, privacy, intellectual property, cybersecurity, and insurance decision.
A New York professional services firm should not wait for a client dispute to discover that:
Its E&O policy excludes a technology service
Its cyber policy excludes professional services
Its media coverage does not apply to the content at issue
Its crime policy has a small social-engineering sublimit
Its business interruption coverage requires physical damage
Its vendor contract disclaims accuracy
Its client contract imposes uninsured indemnity obligations
Its claims-made policy required earlier notice
Insure Connecticut LLC is an independent insurance broker that can help professional services firms review coverage across multiple carriers and identify gaps between E&O, cyber, crime, general liability, property, business interruption, and contractual requirements. Although our company is based in Connecticut, we work with clients across multiple states and can help firms with regional operations involving New York, New Jersey, Massachusetts, and Rhode Island.
Start with our small business insurance resource and 2026 business insurance review. These resources explain how commercial policies work together and why coverage should reflect how your business operates today.
Request a commercial insurance review from Insure Connecticut LLC. Ask specifically for an AI-risk review of your professional liability, cyber, media/IP, crime, general liability, business interruption, and contract requirements.
If your firm has employees, review your workers’ compensation coverage at the same time. Payroll, job duties, classifications, remote work, and multi-state operations can change the insurance program even when your headcount has not changed.
This article is for general educational purposes. It is not legal advice, tax advice, technology advice, or a coverage determination. Insurance coverage depends on the actual policy, endorsements, exclusions, limits, conditions, applicable law, and facts of the claim. Consult qualified legal counsel and your licensed insurance professional for advice about your firm.
Additional references
New York Department of Financial Services official YouTube channel
Small-business practitioner discussions about AI and insurance on Reddit : informal discussion only, not authoritative insurance or legal guidance
.png)

Comments