| Stories | https://www.bloomberg.com/authors/AVvdlCVOfng/davey-alba |
| https://www.twitter.com/daveyalba | |
| Website | https://www.daveyalba.com |
| VERIFIED BY PRESSCHECK.ORG | https://presscheck.org/journalists/davey-alba |
NEW: I wrote about a semi-secret Discord that Google runs for heavy users of Bard.
In it, Google PMs, designers and engineers tasked with developing Bard say they don't know what LLMs are truly useful for, don't trust LLM output before verification, and more
My first (!) Businessweek cover: Investigators are increasingly using warrants to obtain location and search data from Google, even for nonviolent cases—and can end up looking through the user data of people who had nothing to do with a crime.
with my great reporting partner, Julia Love
One of the contractors' guidelines states that reviewing the AI for misleading content should be “based on your current knowledge or quick web search... You do not need to perform a rigorous fact check.”
We're making the documents public here: https://assets.bwbx.io/documents/users/iqjWHBFdfxIU/rqKqEqbXBnDI/v0
NEW: Internal documents shared with me show the complex instructions that Google contractors are asked to complete in minutes in order to train Google’s generative AI products, like Bard.
AI experts say the workers have an "impossible job."
Important reporting from @daveyalba at Bloomberg on the working conditions of the people hired to be "trainers" for Google Bard.
>>
Important reporting from @daveyalba on the impossible job given to annotators working on Google Bard, including amongst other things, verifying critical medical and legal information via quick web searches.
"The trusted internet-search giant is providing low-quality information in a race to keep up with the competition," --- this phrasing makes it starkly clear that it's a race to nowhere good.
From @daveyalba
>>
Leon Yin
Prof. Emily M. Bender(she/her)
Alex Hanna
