Category ChatGPT

Report on the first scientific experiment to test the impact of generative ai on complex knowledge intensive work by Ralph Losey

Report on the First Scientific Experiment to Test the Impact of Generative AI on Complex, Knowledge-Intensive Work

In a groundbreaking study involving almost 250 consultants from Boston Consulting Group, conducted by Harvard, Wharton, Warwick, MIT, and BCG, revealed significant enhancements in task quality and productivity through AI integration. The experiment, detailed in Harvard Business School's Working Paper, showed a striking almost 40% increase in task quality and a 12% productivity boost when using AI tools like ChatGPT4. However, tasks beyond AI's current capability resulted in a decline in correct outcomes, emphasizing the need for appropriate task selection. The research underscores AI's potential as a "skill leveler" among professionals, reshaping work dynamics.
Read MoreReport on the First Scientific Experiment to Test the Impact of Generative AI on Complex, Knowledge-Intensive Work
Centaur with cyborg head and torso, with cape in the wind

From Centaurs To Cyborgs: Our evolving relationship with generative AI

[Editor’s Note: EDRM is proud to publish Ralph Losey’s advocacy and analysis. The opinions and positions are Ralph Losey’s copyrighted work.] Centaurs are mythological creatures with a human’s upper body, and a horse’s lower body. They symbolize a union of…

Read MoreFrom Centaurs To Cyborgs: Our evolving relationship with generative AI
Stochastic Parrots: the hidden bias of large language model AI by Ralph Losey.

Stochastic Parrots: the hidden bias of large language model AI

The subtle biases of GPTs can be an even greater danger than the more obvious problems of AI errors and hallucinations.  We need to improve the diversity of the underlying training data,  the  curation of the data, and the Reinforcement Learning from Human Feedback, RLHF. It is not enough to just keep adding more and more data, as some contend. This view was forcefully argued in 2021 in an article I recommend.  On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? (FAccT ’21, 3/1/21) by AI ethics experts, Emily M. Bender, Timnit Gebru, Angelina McMillan-Major and Margaret Mitchell.
Read MoreStochastic Parrots: the hidden bias of large language model AI
OpenAI Generates a 'Hired Gun Hacker' Defense to the N.Y. Times Copyright Case by Ra;ph Losey

OpenAI Generates a ‘Hired Gun Hacker’ Defense to the N.Y. Times Copyright Case

The NYT alleges copyright infringement by the practice of OAI, and most all other generative AI companies, of using data scraped from the internet to train their generative AIs. The defendants responded in late February with motions to dismiss. OAI's motion raised a novel "hired gun hacker" defense that intrigued AI hacker attorney, Ralph Losey. Here is Ralph's report on the defense,
Read MoreOpenAI Generates a ‘Hired Gun Hacker’ Defense to the N.Y. Times Copyright Case
EDRM’s Upcoming Workshop: A Deep Dive into Generative AI and eDiscovery by Rob Robinson.

EDRM’s Upcoming Workshop: A Deep Dive into Generative AI and eDiscovery

Heads up, legal tech enthusiasts! The EDRM is rolling out something you won’t want to miss on March 19: a deep-dive workshop all about Generative AI and its game-changing role in eDiscovery. Forget the usual dry and technical spiel; this is your ticket to understanding how AI is really shaking things up in our world. With a superstar lineup featuring the likes of Dr. Maura R. Grossman, John Brewer, and David R. Cohen, plus EDRM’s own Mary Mack and Kaylee Walstad, steered by the ever-engaging Mary Bennett, we’re in for a real treat.
Read MoreEDRM’s Upcoming Workshop: A Deep Dive into Generative AI and eDiscovery