The use of Artificial Intelligence in the legal industry continues to be an evolving and highly discussed topic. While you may be getting updates from other sources on news and developments, we wanted to synthesize a few important developments on the generative AI front.
1. Lawyers sanctioned for sloppy use of ChatGPT and reaction from the judiciary.
In our May post about AI Technology, we cautioned attorneys that current Learning/Large Language Models such as ChatGPT are incapable of conducting legal research. These systems can “hallucinate” caselaw, statutes, or other sources because they are capable of aggregating data and reconstituting it into something that looks believable.
Unfortunately, some New York attorneys hadn’t gotten that warning prior to filing a March 1 brief which included citing to fake cases. When challenged on it, they didn’t immediately acknowledge the error, even filing an affidavit nearly 2 months later that “doubled down” by attaching copies of fabricated excerpts. At the sanctions hearing, the attorneys then testified that they were “operating under the false perception that [ChatGPT] could not possibly be fabricating cases on its own” and that the application was a “super search engine.” Predominantly due to the attorneys’ avoidance of responsibility after being alerted to the error, the federal judge found that the attorneys acted in bad faith and with “conscious avoidance and false and misleading statements to the court.” As sanctions, the judge ordered the attorneys, jointly with the firm, to send a letter to each judge falsely identified as the author of the fake opinions along with a copy of the sanctions order and fake opinion, and to pay a penalty of $5,000, along with other required acts.
With the New York matter pending, a federal judge in Texas issued an order requiring attorneys to include a certificate with filings to indicate either that no portion of any document they file was created by generative AI, or if it was, that a human being had reviewed and checked any AI-generated text. In justification, Judge Brantley Starr wrote:
These platforms in their current states are prone to hallucinations and bias. On hallucinations, they make stuff up—even quotes and citations. Another issue is reliability or bias. While attorneys swear an oath to set aside their personal prejudices, biases, and beliefs to faithfully uphold the law and represent their clients, generative artificial intelligence is the product of programming devised by humans who did not have to swear such an oath. As such, these systems hold no allegiance to any client, the rule of law, or the laws and Constitution of the United States (or, as addressed above, the truth). Unbound by any sense of duty, honor, or justice, such programs act according to computer code rather than conviction, based on programming rather than principle.
2. OpenAI is being sued – in a big way.
Authors Paul Tremblay and Mona Awad recently brought a class action lawsuit against OpenAI (the parent company of ChatGPT) alleging that the programs have been trained on copyrighted works without consent, attribution, or compensation. The complaint alleges that OpenAI has admitted to training GPT on copyrighted literature through use of “internet-based books corpora.” Due to its training on copyrighted material, including works written by the named plaintiffs, the complaint alleges causes of action for copyright infringement, unfair competition, negligence, and unjust enrichment. In a separate class action, OpenAI was sued for unlawfully collecting personal information in violation of federal and state privacy laws. Termed “data scraping,” the lawsuit alleges that OpenAI extracts data from websites into its own files, databases or applications. In Plaintiffs PM, et.al v Open AI LP, et. al, 3:23-cv-03199, filed in the Northern District of California, the plaintiffs argue that the platform creates an “existential threat” and links “unchecked AI proliferation” exemplified by OpenAI to autonomous weapons, among other catastrophic threats to society.
With both complaints just being filed at the end of June, outcomes are likely years away, but they do give the judiciary an opportunity to hold the reigns to an extent on AI development, perhaps while legislation is being explored.
3. Legal-sector specific generative AI in beta testing.
It’s not all dark and foreboding on the generative AI horizon. Lexis+ AI is in beta testing and Thompson Reuters is promising “generative AI capabilities in Westlaw Precision” by the end of 2023. These closed-system generative AI tools tout apps that use proprietary caselaw databases. Lexis+ AI claims that the program cannot hallucinate cases and will produce complete and accurate summaries. There’s a lot of optimism that the transition from analytic AI already in use to generative AI will be as transformative as the move to online legal research in decades’ past.
If all of this still sounds like a new language, here are some resources:
Ohio State Bar Association CLE webinars:
The Rise of AI in the Legal Profession: Lawyers Brace for Impact (Live Webcast on 07/24/2023)
Tech Never Stops (On Demand through 12/31/2023)
The State of Artificial Intelligence in Recruiting (On Demand through 12/31/2023)
Proskauer Rose Age of AI Webinar Series (Complimentary, not pre-approved for Ohio CLE credit)
As we wrap up this week’s post, we agree that Artificial Intelligence in its various forms is here to stay and having a huge impact on the legal industry. If you remember the shift from analog to digital, we are perhaps on the crest of a similar wave. But at OBLIC, we’re here to help you stay afloat, even if you’d rather not surf in the AI tide.
|Gretchen K. Mote, Esq.
Director of Loss Prevention
Ohio Bar Liability Insurance Co.
|Merisa K. Bowers, Esq.
Loss Prevention Counsel
Ohio Bar Liability Insurance Co.
This information is made available solely for loss prevention purposes, which may include claim prevention techniques designed to minimize the likelihood of incurring a claim for legal malpractice. This information does not establish, report, or create the standard of care for attorneys. The material is not a complete analysis of the topic and should not be construed as providing legal advice. Please conduct your own appropriate legal research in this area. If you have questions about this email’s content and are an OBLIC policyholder, please contact us using the information above.