How to Safeguard Human Judgment in Legal AI Adoption

In-house adoption of generative and agentic AI continues to grow rapidly. Some 67% of in-house professionals are now using generative AI at work, according to data from a recent Association of Corporate Counsel (ACC) survey. Many have adopted AI to help them draft everything from summaries and communications to full legal agreements, while an increasing number deploy “legal agents” to handle complex, multistep workflows. As agentic AI advances, the critical challenge is helping attorneys effectively manage risk, maintain control of their work, and ensure that final documents and decisions reflect their own human judgment.

While professionals use AI to create consistent work product based on company standards, legal leaders recognize the need to provide governance and training that keeps pace with practice. The most successful teams keep human legal judgment at the center. Your oversight can help attorneys make informed decisions, stay compliant, and avoid the risk of court sanctions.

What you need to know about risk

Even now that generative AI has become a household word, court sanctions for attorney misuse of AI are still prominent in the headlines. Recent articles have reported sanctions over fake case quotes, misrepresentations of the law, generative AI citation hallucinations, and undisclosed or poorly explained uses of AI. These are actions that judges may categorize as abuse of the judicial process.

Clearly, it’s critical to minimize risk that could stem from unverified outputs in filings, advice, discovery responses, board materials, or other legal work. This responsibility goes beyond policies and reminders to validate quotes and citations.

Attorneys need to understand what types of information are acceptable to provide to an AI system, how to validate work, and what documentation is needed to demonstrate exercise of professional judgment. The established process must be something they can explain, replicate, and defend. Making this happen is a matter of leadership and a commitment to best practices.

Implementing best practices

Making sure your team avoids risk and builds success requires more than broad AI policies and training. Here’s quick list of actions to consider:

• Collaborate on in-house training. Assign a trusted team member to take the lead and serve as a resource for colleagues. That in-house training leader should also have an active role on your company’s broader AI governance committee, if possible.
• Make sure training sessions and follow-up conversations underscore the pros, cons, and risks of AI in legal tasks likely to arise. Remind lawyers about ways their competence, confidentiality, supervision, accuracy, vendor due diligence, and other obligations apply to AI in day-to-day work.
• Make sure your approach complies with applicable federal and state laws, regulations, and regulatory guidance on development or deployment of AI, impacts on individuals (e.g., consumer protection, privacy), and industry-specific requirements.
• Explore legal-specific AI platforms that can help your team stick with company-approved workflows, trusted legal content, and organizational standards. Consider an offering that combines legal skills, workflows, and validation tools to help professionals expedite work while maintaining control. For complex, multistep, or higher-judgment legal work that can benefit from agentic AI, ask about tools that enable users to review and approve a proposed, skill-based approach before execution begins. When exploring free trials and selecting vendors, be sure to check your company’s AI and purchasing policies.

Resources to take you forward

By working with your training leader, you can better prepare and motivate your team to apply AI best practices. Here are some resources to support that work:

• The ACC Artificial Intelligence Toolkit for In-house Lawyer, Second Edition, offers practical checklists for developing AI governance, addressing third-party risk, meeting ethical obligations, and evaluating agentic AI oversight. You’ll find tips for conducting an AI impact assessment, protecting intellectual property, and addressing governance for agentic AI. ACC also provides The AI Playbook: Putting AI to Work for In-House Legal Teams, which emphasizes the importance of targeting defined legal workflows rather than adopting AI as a general-purpose solution.
• Industry-specific and educational organizations provide AI oversight guidance and certifications, such as the Joint Commission’s certification for Responsible Use of AI in Healthcare and the International Association of Privacy Professionals credential for artificial intelligence governance professionals.
• The American Bar Association offers insight on AI literacy for in-house attorneys, including this recent recorded webinar entitled “In-House Insights: AI Literacy 101 for In-House Counsel.”
• CLE training is available from professional organizations and vendors on topics like AI in employment, drafting with generative AI, and ethical considerations involving use of technology in legal practice.
• Your legal operations or purchasing team can help you closely examine the capabilities of in-house legal AI tools. Ask vendors about purpose-built, legal AI-enabled surveys of laws and regulations. See if outputs are grounded in vetted, attorney-authored content providing insight on topics like AI-enabled threats and mitigation strategies, as well as compliance and risk management issues. Ask vendors about the grounding behind any AI legislation and regulation trackers.

Reasons for confidence

Using the tips and resources outlined above, you can orchestrate a generative and agentic AI approach that supports professional judgment and control. As your team explores and adopts advancements in AI, your exercise of leadership will help them expedite work, reduce risk, and support the opportunities that matter most to your business.