Lots of people are talking about ChatGPT.
Some, like those at Microsoft, see it as a valuable tool to be integrated into their products and platforms; indeed, one of its lawyers thought that the answer provided by ChatGPT to his legal question “sounds like a pretty good lawyer.” But others are wondering whether we should be “terrified” of ChatGPT, and even those knowledgeable about computer programming can be “freaked out” by the program. Some are fueling the terror by comparing ChatGPT’s release to the unleashing of the Kraken. And the debate continues, with Seattle, Los Angeles, NYC, Fairfax County (Virginia), and Montgomery County (Alabama) for example, having banned its use on or blocked its access on their public school systems and devices, while other educators are advocating its incorporation into the curriculum, according to both Forbes and the New York Times.
Comparing the content-generating AI platform to the Kraken makes sense. “’Kraken’ comes from the German word for ‘octopus,’” Like an octopus, ChatGPT seems to be reaching out into multiple different directions simultaneously; at least for now, the best one can say about ChatGPT replacing students, teachers or lawyers is that “[t]hese are certainly mythical tales,” and can either be viewed as “typically rely[ing] on a grain of truth that gets distorted” or as helpful explanations “based in fact.” While, for some, the release of ChatGPT seems accompanied by a dramatic feel, my experience with ChatGPT started with a much more sitcom feel but had an admittedly surprising ending.
Why? Because I thought ChatGPT might get me off to a great start on this ILN article. I had last time out expressed an interest in following developments in Jack Daniels v. VIP Products, the so-called “Bad Spaniels” case. The slew of amicus and other briefs had been filed, and it was time to “get crackin’ on” this blog post, as in time to get moving on the drafting. So, I signed up into ChatGPT, and entered the query “compare and contrast arguments of amicus curiae in ‘bad spaniels’ case,” only to get the response that ChatGPT was unfamiliar with that case, so I entered the formal name and docket information for the case, and got a serviceable summary of the parties’ positions in the case. But as to the positions of amici, all I got was “In amicus briefs filed in the case, some parties likely supported the arguments put forth by Jack Daniel’s, while others likely supported the arguments of VIP Products.” (Emphasis added) Not much help, and when I added some of the names of specific filers of such briefs, I did not get a summary of their arguments, much less contrast against others. I got a statement noting that given the name of the organization entered, it is “likely” that such an amicus “would have supported Jack Daniel’s argument” and “[t]hey may also have added” certain things to the argument. This left me still needing to get crackin’ on the assignment in that first sense of moving forward and wanting to simply be cracking on (i.e. making fun of) the weakness of the ChatGPT response. So, in this post, I guess I am doing both.