Introduction:
In a world undergoing massive change due to the development and increasing use of artificial intelligence (“AI”), the role of AI in providing legal services has become a topic of increasing complexity and concern. Early AI was poorly used by lawyers with AI literally fabricating case law and case history resulting in a furious court that sanctioned the attorneys who utilized the AI for misrepresenting the law to the court. Early AI has been improved even if it is still far from fool proof and its efficiency and cost benefit has forced the legal profession to utilize it more commonly.
Indeed, as discussed in this article, it may be ethically required for California business lawyers to become conversant and skilled at use of AI. Traditionally a conservative profession, the remarkable efficiency and cost cutting that AI provides is forcing attorneys to adjust to this new reality.
This article shall discuss the state of the law both on use of AI and the dangers to both the legal profession and public of inappropriate use of AI.
Basic Issues of Utilizing AI:
From the point of view of the courts, AI is simply a tool, not legally different from a typewriter or dictating machine. Thus, the human attorneys are ultimately responsible for use and misuse of the work product of AI. Put simply, the attorney is liable for errors whether known or unknown caused by AI or caused by his or her misuse of AI.
The types of errors attorneys can commit can be classified as follows:
- The Risk of Overreliance. AI always sounds confident and in the high-tension world of providing legal services where each minute can cost fifteen dollars, it is tempting to rely on the generated data provided by AI and not spend the time and effort to check all the data provided. The most common source of AI-related malpractice and sanctions involves blindly trusting generative AI output.
- The “Hallucination” Illusion: Consumer-grade AI tools can often fabricate non-existent legal precedents or misapply citations. Submitting these to a court violates the attorney’s duty of candor. Note that later versions of AI, costing a good deal more than the typical consumer AI, is far less likely to make this type of error. That said, care is required. It is the lawyer’s license on the line.
- Delegating Legal Judgment: The American Bar Association (ABA) mandates that lawyers cannot delegate professional judgment to any unlicensed person…or machine. Failing to perform human verification of AI-generated work constitutes professional negligence.
But lawyers cannot just elect to ignore AI to avoid the above risks. Due to the evolving legal ethical standard of care, it may be required to utilize AI. Many legal experts and insurance professionals are now identifying an entirely new malpractice exposure: the failure to use AI. Failing to utilize available technological tools to competently and efficiently represent a client could eventually be argued as a breach of the standard of care.
Further as AI-assisted workflows become an industry norm, clients increasingly demand the efficiency and cost-reductions that AI provides. A standard contract, as more fully described below, can normally be created in seconds by AI while an attorney, especially an inexperienced one, could take multiple hours at many hundreds of dollars an hour. How does the attorney bill that sum for a finished product that with AI took three seconds to produce and requires perhaps twenty minutes to proof?
Ethical Frameworks for AI Generated Services:
The American Bar Association and various state bar associations (such as the State Bar of California) have issued formal guidance requiring lawyers to adhere to the following when working with AI:
- Duty of Competence: Lawyers must understand the benefits, risks, and limitations of the AI tools they use.
- Confidentiality: Entering sensitive, non-public client data into public or unsecure AI models can waive attorney-client privilege and cause data breaches.
- Supervision: Attorneys are ultimately responsible for the work product, whether generated by an associate, a paralegal, or an algorithm. Conversely, overreliance on AI without adequate verification may violate duties of diligence, candor, and supervision.
These guidelines of the State are binding on attorneys practicing in the state and failure to comply also is powerful evidence of malpractice by any client that is injured.
AI and Contract Law
AI is uniquely suited to drafting and negotiating contracts. Most contracts have “boiler plate” provisions and even when the average California Contract Lawyer drafts without use of AI, that attorney is likely to have provisions in memory that he or she cuts and pastes into the standard contract. AI would seem to be ideal to utilize in such repetitive situations.
But while AI automates drafting, summarization, and risk identification, under Bar rules it does not replace human judgment.
The typical contractual areas AI is useful for are as follows:
- Lawyers and businesses use AI to draft routine contract tasks, increasing speed and accuracy.
- Drafting & Redlining: AI can instantly draft agreements, spot deviations from standard playbooks, and suggest revisions based on past successful negotiations.
- Review & Due Diligence: Tools like Kira Systems or Spellbook can scan thousands of contracts in seconds to extract specific clauses, summarize complex terms, and identify hidden risks.
- Lifecycle Management: AI integrated into Contract Lifecycle Management (CLM) platforms automate renewal tracking, tracks compliance, and manages approval workflows.
Even in the realm of contractual arbitration or litigation, AI can be a powerful tool.
- Legal Interpretation & Research: AI is reshaping how the plain meaning of contracts is evaluated.
- Contextual Analysis: Advanced Large Language Models (LLMs) help lawyers and judges assess the meaning of ambiguous clauses by quickly analyzing vast datasets of past policies and rulings.
- Consistency: AI ensures standard definitions and clauses are used uniformly across an organization’s entire document repository, reducing human error.
- Autonomous Agents: The rise of B2B AI agents capable of autonomously negotiating and making contracts is challenging traditional legal principles of “meeting of the minds” and intent. As an example which demonstrates future trends, the agreements governing the use of AI software (e.g., terms of service and vendor contracts) are actively dictating AI governance, data ownership, and liability allocation.
Danger and Practicality:
From the legal consumer’s point of view, the issue is not complex. The attorney remains civilly and ethically responsible for all actions of AI the attorney has used. This remains true whether or not the attorney was negligent in the use of the AI or the AI, itself, had a flaw which the attorney did not know. For the attorney to blame the AI and seek to avoid liability is akin to the attorney blaming his or her pencil for writing the wrong contract.
For the attorney the issue is both more complex and more troubling. While responsible for anything that AI does, the attorney is required to utilize AI as a technological requirement for the efficient and appropriate practice of law. The attorney is thus required to learn and master use of AI, and this will be required for all aspects of the practice of law, with checking AI generated work of any kind carefully and thoroughly.
