I attended the American Bar Association’s Cyber & Technology Committee’s Winter Working Meeting last week in Fort Lauderdale, Florida, where we caught up on the latest developments in the Cyber Technology space. Needless to say, the majority of the conversations revolved around artificial intelligence (AI). Following are my key take-aways:
- Canada is Going Prorogue
So it turns out Canada has a unique law that allows the Prime Minister to suspend all parliamentary activity. Justin Trudeau did just that in early January of this year. This means that all laws that had been proposed, including new and innovative privacy and AI laws favored by the more liberal administration have been suspended. The suspension of the privacy laws in particular could be indefinite depending on whether a conservative political faction is elected in mid-March of this year.
In other Canadian developments, the nation has seen an unprecedented uptick in case law. For instance, in Zhang v. Cheng, a British Columbia lawyer filed a notice of application containing non-existent legal authorities that had been “hallucinated.” The court determined that the lawyer was legally responsible for the hallucinations of the AI tool she used. This case highlights the ethical obligations of lawyers when using AI tools, including the duty of competence and the need to verify AI-generated outputs.
Continuing with the AI liability theme, the Moffat v. Air Canada case is a significant decision regarding the liability of companies for information provided by AI chatbots. The airline carrier was found liable for the misinformation its chatbot generated. At court, Air Canada tried to argue that the Chatbot was a separate “person,” and as such the carrier should not be responsible for the misinformation provided. The court did not buy that argument. This case underscores the importance of companies taking reasonable care to ensure the accuracy of information provided by AI chatbots and highlights the potential legal liabilities associated with AI-generated content.
- Walled Gardens Are Still Hallucinating
“Walled Gardens” is the phrase used to describe a closed ecosystem or platform that restricts the flow of information and services to and from external sources. According to Iterate.AI “[i]t is like a controlled environment where access is limited and tightly regulated. In the context of technology, it often refers to a company or platform that tightly controls what users can access and do within its ecosystem, often at the expense of interoperability with external services.” Unfortunately, despite efforts to fend off bad data in the AI space, walled gardens are still susceptible to hallucinations. The percentage of hallucination is much smaller than in the open platforms, but still the percentage is in the high teens which is not acceptable.
- Maryland’s New Privacy Law Prohibits Sale of Sensitive Personal Information
Maryland’s new Online Data Privacy Act, which will become effective in October of 2025, is designed to align with the omnibus privacy laws already enacted in other states. This law is critically different because it prohibits the sale of sensitive personal information. This means that even with consent, a provider cannot in any way sell or exchange sensitive personal information for any sort of consideration. The concern here is not on monetary sale, but in the act of exchanging information for consideration which is much trickier to navigate.
- 30 Long Pending Cases on AI and Copyright Issues Still Unresolved
There are 30 cases dealing with the thorny subject of AI and copyright issues. Unfortunately none have come to a conclusion except for Raw Story Media v. OpenAI, which was dismissed but will likely be refiled. Perhaps the most watched case is the Getty Images v. Stability AI (D. Delaware) case. Here the Plaintiff claims Stability AI infringed on more than 12 million of its photographs, their associated captions and metadata. Unfortunately, all parties are currently in jurisdictional limbo related to the Defendant’s motion to transfer the case. Therefore, no substantive progress to report.
Similarly, in New York Times v. Microsoft (S.D.N.Y.) the NYT alleges that Microsoft used millions of its copyrighted works to create the LLM’s of Co-Pilot (formerly Bing Chat) and OpenAI’s ChatGPT. The NYT asserts that these AI tools generate verbatim NYT content, closely summarize it, mimic its expressive style, and falsely attribute outputs to the NYT. The arguments are indeed fascinating, but it may be a while before any substantive decisions are published.
- Various AI Laws Are Now DOA
At least one speaker at the conference was very pessimistic (some might say optimistic) about the attempted regulation of AI. For instance, California’s SB-1047 seeking to control “safe and secure innovation for Frontier AI Models” is now defunct. Similarly, he is predicting that Colorado’s much ballyhooed AI Act will soon meet a shallow grave. This is because the Colorado AI law has a narrow focus on high-risk AI systems, which are already the subject of regulations that the new U.S. executive administration is seeking to curtail.
On the whole, tensions and concern continue in the AI space. Although there appears to be a short-term momentum towards easing restrictions, the private sector is still looking for guardrails and may even impose self-regulation. Besides that, state governors are continuously reminding anyone who will listen that their omnibus privacy laws indeed apply to AI. So, whether you are directly in the AI space or simply processing personal information, none of this means you can let your legal compliance guard down.
The information you obtain at this site, or this blog is not, nor is it intended to be, legal advice. You should consult an attorney for advice regarding your individual situation. We invite you to contact us through the website, email, phone, or through LinkedIn. Contacting us does not create an attorney-client relationship. Please do not send any confidential information to us until such time as an attorney-client relationship has been established.