Overcoming the black box

In looking at data from the Bureau of Justice Statistics (part of the US Department of Justice), Boston University has some questions about bias in AI used in legal. The post is a great reminder that for any good development project, making sure you have input from all the constituency involved and impacted. This is just more reasons why legal must overcome the black box and take the lead in transparent, explainable AI. Read more at REAL KM: Paper highlights the bias inherent in legal AI

- o/|oo|\o -

They're ba-ack!

Remember Ross? Not the guy with the monkey but rather the former legal AI research company. They are back in the news. Johnny-on-the-spot, Bob Ambrogi, reports that 3rd US Circuit Court of Appeals Judge Stephanos Bibas ruled against both parties motions for summary judgment, sending the case to trial for copyright infringement. But as Bob notes, "However, on one critical issue, Bibas ruled in favor of Thomson Reuters, finding 'as a matter of law' that Ross actually copied Westlaw headnotes and other TR copyrighted content." Why send it off to trial? The judge felt "the case is a 'factual mess' that needs to be sorted out by a jury after trial, not by a judge on summary judgment." Read more at LawSites: Breaking: Judge Rules Ross Intelligence Copied Westlaw's Headnotes, But Leaves for Jury to Decide if It Violated Thomson Reuters' Copyright

- o/|oo|\o -

Trust nothing

Surfing the web the other day I saw a meme about Star Trek predicting our technology future. It started with Kirk and the 1966 TOS era communicator predicting the flip phone of 1973 and ended with 1988 DS9 Sisko predicting wearable eye tech in 2013. Now it seems we can add to that meme Uhuru and the 1966 TOS era universal translator (or maybe that should be the 2001 Enterprise era Hoshi Sato? Prequels always mess me up.) predicting 2023 OpenAI/Spotify's blog translator. No way tech that can translate your own voice to other languages can be abused! And the second highlight discusses that ChatGPT 3.5 and 4 are being enhanced with the ability to "analyze images and react to them as part of a text conversation" and the mobile apps will "enable fully verbal conversations." I've got new roles for two of the three wise monkeys. "Trust nothing that is seen. Trust nothing that is said." Read more at:
Silicon: Spotify AI Tool Translates Podcasts In Speaker's Own Voice
ars technica: ChatGPT update enables its AI to "see, hear, and speak," according to OpenAI

- o/|oo|\o -

Jeffrey Brandt, Editor

Connect with me on LinkedIn Jeffrey Brandt
Follow me on Twitter @jeffrey_brandt
Follow all the PinHawk highlights @PinHawkHappens
Website: PinHawk.com

Affiliations:
ILTA Strategic Partner Liaisons - NetDocuments