Recently, the use of generative artificial intelligence (AI) in Right-to-Know Law (RTKL) matters has become increasingly commonplace. While nothing currently prohibits the use of AI when crafting a request, appeal, or submission to the Office of Open Records (OOR), users should be aware of the risks before doing so.
AI chatbots are currently not reliable tools for preparing submissions under the RTKL and should not be considered a substitute for legal counsel or professional representation. These tools frequently generate inaccurate or misleading information, including fabricated legal citations, erroneous summaries of case law, and false quotations from court decisions and OOR Final Determinations.
Such errors—commonly referred to as “AI hallucinations”—involve the creation of content that may appear authentic but is factually incorrect or entirely fictitious. Filings to the OOR have included citations to non-existent cases, mischaracterizations of legal precedent, erroneous summaries of case law, and invented language attributed to legal authorities. Parties using AI-generated content do so at their own risk and remain responsible for ensuring the accuracy of anything they submit during an OOR appeal. Parties should never rely upon AI for legal advice.
The submission of AI-generated content without independently verifying the accuracy of its statements may subject parties to sanctions. Repeated submission of unverified AI-generated material—especially after prior inaccuracies have been identified—may result in additional penalties. Under the RTKL, courts have the authority to impose sanctions, including fines and attorney’s fees, for frivolous arguments. See 65 P.S. § 67.1304(b), (c). Knowingly submitting false information to the Office of Open Records (OOR) or a court may also carry criminal consequences under Pennsylvania law. See 18 Pa.C.S. § 4904(a). Finally, OOR Appeals Officers have the discretion to reject any filing that includes false statements, misrepresentations, fictitious legal citations, or duplicative content generated by AI.
While AI may be useful in some contexts, it currently struggles in the RTKL setting. Please do not submit AI-generated content to the OOR without first checking its accuracy. Parties are encouraged to point out false statements, nonexistent legal precedent, and misrepresentations, whether AI-generated or otherwise, to their assigned Appeals Officer.
When processing and responding to RTKL requests, Agency Open-Records Officers (AORO) should be aware that they may be presented with nonexistent or misleading legal precedent that would appear to require the release of information that is otherwise exempt or confidential. AOROs should direct any questions about the authenticity of legal precedent to their solicitor(s).
What about AI being the subject of RTK requests?
LikeLike
Any records request should be reviewed on a case-by-case basis. We are unaware of any specific rulings on records related to AI.
LikeLike