1. SPY FEDS BUILDING AN AI CHATBOT
The whole bridge of spies in the U.S. is getting its own ChatGPT-style AI. This week, the Central Intelligence Agency confirms it’s building a bot the entire 18-agency intelligence community will use, including the CIA, NSA, FBI, and military offices. The Agency’s director says the public data generated just keeps going and they need a means of finding the eye of the needle in a field of needles. The bot will reportedly allow agents to tap information and ask follow-up questions and summarize masses of data. CIA leaders say with AI they can grow their information collection with no limitations other than cost. Not concerning at all.
2. FEELING THE BUNSEN BURN OF BAD SCIENCE
True crime’s got online sleuths and academia has data debunkers. A Harvard professor’s academic papers on why people cheat numbered among the 5,500 studies retracted in 2022. So did three studies from Stanford’s (former) president. That’s up from 119 in 2002. Weird science? No – increased scrutiny from scientists moonlighting as scholarly-journal police looking for fraud. Experts say academics face intense pressure to publish papers yielding unique findings, which can generate jobs, grants, and prestige. But data debunkers say they’ve got to keep science honest. Public faith is already shaky – we see you flat earthers – and flawed research wastes years and money while inspiring faulty corporate decisions or misguided policies.
3. JOURNALIST NAMED BABY “METH RULES” FOR STORY
An Australian journalist, nine months pregnant, was working a story about what parents can legally name their newborns. The registrar of births theoretically denies offensive or obscene names. But she wanted to know what default names officials bestow in their place. With contractions less than six minutes apart, no answer from the registry, she felt a need for speed, so she birthed her baby and named him Methamphetamine Rules. Then she waited for the Registrar to reject his name and tweak it. But Meth Rules, slipped through the crack. Mom didn’t realize it until his birth certificate arrived. The registrar says they’ve since cranked up their monitoring.
4. AUTHORS SUE OPENAI IN REAL CLIFFHANGER
A Group of prominent authors are pursuing a lawsuit that could re-write the book on Artificial Intelligence. Plaintiffs, including George R.R. Martin and Jodi Picoult, accuse OpenAI, creator of ChatGPT, of scanning and using their copyrighted work without permission. The case targets generative AI’s massive ingestion of billions of pieces of text and authored works as training fodder, from which it instantly composes poems, music, computer code and much more when prompted. Experts say fact, not fiction, if anyone’s going to win on copyright-infringement claims against OpenAI – it’s these lead characters. But there are possible plot twists. OpenAI could argue “fair use” defense, saying what ChatGPT generates is transformative.
5. DNA TESTING FOR PETS? GET A POOPER SCOOPER
Humans digging for the genetic scoop on their furry friends have bred a $345M Pet DNA industry. The kits fetch up to $80 a piece, but experts say they retrieve, at best, spotty results. One consumer tested her German Shepherd-resembling rescue pooch. Results showed shepherd, but also Chihuahua. Other pet DNA companies labeled her dog part Siberian Husky and Great Pyrenees. When the owner sent in samples from her human cheeks, tests showed her to be nothing but a hound dog – well, actually 40% Border Collie and 28% Bulldog. Experts say the science is fuzzy. Breeds are based primarily on a dog’s physical looks, not the appearance of their genes.