Search results

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2xfo @RnDanger@infosec.exchange
Public

I'm seeing people upset at Fedora for allowing AI assisted code submissions, but I'm seeing it like people are mad that AI will access their data and i didn't think that's anywhere close to what's happening.

Fedora is letting coders use AI as long as they admit it was assisted code and take "responsibility" for it. They aren't adding bots to the system or taking screenshots for analysis like Windows is doing.

I think I'm pretty anti-AI but this feels mostly ok to me

0
0
0
0
0
0
0
0
0
MidsouthMouth.OctaviaKeats @fkaOctaviaKeats@wandering.shop
Public

Tell me, if *I* regularly "hallucinated" on the job, how long before I was separated?
AI up here being treated like a rich white man's nephew in the workplace!

"Researchers say AI transcription tool used in hospitals invents things no one ever said" from AP News







apnews.com/article/ai-artifici

Assistant professor of information science Allison Koenecke, an author of a recent study that found hallucinations in a speech-to-text transcription tool, works in her office at Cornell University in Ithaca, N.Y., Friday, Feb. 2, 2024. The text preceded by "#Ground truth" shows what was actually said while the sentences preceded by ""text"" was how the transcription program interpreted the words. (AP Photo/Seth Wenig)

Researchers say an AI-powered transcription tool used in hospitals invents things no one ever said

Whisper is a popular transcription tool powered by artificial intelligence, but it has a major flaw. It makes things up that were never said.

apnews.com · AP News

0
0
0
Miguel Afonso Caetano @remixtures@tldr.nettime.org
Public

"Google AI overviews are misleading or inaccurate in 37% of finance-related searches, according to The College Investor's latest analysis. This is an improvement from last year, where 43% of AI Overviews were inaccurate - but a one-third error rate is troubling when it comes to personal finance.

This is causing consumer confusion, and potentially harming Americans' finances. The overviews were especially bad when it comes to tax, insurance, and financial aid related queries.

What's Happening: Over the last several years, Google has been rolling out AI-driven answers in search results. At the top of the search results they show AI Overviews, and they're now expanding the use of AI Mode. The problem is they are plagued with inaccurate answers. And experts say it's a serious issue."

thecollegeinvestor.com/66208/3

We Asked Google Search 100 Money Questions - Here’s What It Got Wrong

Google’s AI Overviews still struggles with personal finance topics, from student loans to taxes, according to new research.

thecollegeinvestor.com · The College Investor

Link author: The College Investor@thecollegeinvestor@mastodon.social

0
0
0
0
0
0
José A. Alonso @Jose_A_Alonso@mathstodon.xyz
Public

Mathematics with large language models as provers and verifiers. ~ Hieu Le Duc, Leo Liberti. arxiv.org/abs/2510.12829

arXiv logo

Mathematics with large language models as provers and verifiers

During 2024 and 2025 the discussion about the theorem-proving capabilities of large language models started reporting interesting success stories, mostly to do with difficult exercises (such as problems from the International Mathematical Olympiad), but also with conjectures [Feldman & Karbasi, arXiv:2509.18383v1] formulated for the purpose of verifying whether the artificial intelligence could prove it. In this paper we report a theorem proving feat achieved by ChatGPT by using a protocol involving different prover and verifier instances of the gpt-5 model working collaboratively. To make sure that the produced proofs do not suffer from hallucinations, the final proof is formally verified by the lean proof assistant, and the conformance of premises and conclusion of the lean code is verified by a human. Our methodology is by no means complete or exact. It was nonetheless able to solve five out of six 2025 IMO problems, and close about a third of the sixty-six number theory conjectures in [Cohen, Journal of Integer Sequences, 2025].

arxiv.org · arXiv.org

0
0
0
0
0
0
0
0
0
Kevin Marks @KevinMarks@xoxo.zone
Public

Key findings:
45% of all AI answers had at least one significant issue.
31% of responses showed serious sourcing problems – missing, misleading, or incorrect attributions.
20% contained major accuracy issues, including hallucinated details and outdated information.
Gemini performed worst with significant issues in 76% of responses, more than double the other assistants, largely due to its poor sourcing performance.

bbc.co.uk/mediacentre/2025/new

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory

An intensive international study was coordinated by the European Broadcasting Union (EBU) and led by the BBC

www.bbc.co.uk

0
0
0
leyrer @leyrer@23.social
Public

netzpolitik.org: Wie die öffentliche Verwaltung mit KI noch abhängiger von Big Tech wird
netzpolitik.org/2025/sap-und-o

Tirol: Laut der Landesregierung soll die Verwaltung durch KI beschleunigt und vereinfacht werden.
tirol.orf.at/stories/3326785/

Salzburg: Salzburgs Gemeinden wollen mehr Künstliche Intelligenz (KI) in der Verwaltung einsetzen.
salzburg.orf.at/stories/332635

SAP und OpenAI: Wie die öffentliche Verwaltung mit KI noch abhängiger von Big Tech wird

Künstliche Intelligenz soll die Digitalisierung der öffentlichen Verwaltung beschleunigen, so die Hoffnung der Bundesregierung. Passende Angebote kündigen SAP und OpenAI für kommendes Jahr an. Doch damit würden sich Behörden von Bund, Ländern und Kommunen noch enger an die Produkte der Tech-Konzerne binden.

netzpolitik.org · netzpolitik.org

0
0
0
0
0
0
0
0
0
knoppix @knoppix95@mastodon.social
Public

Amazon’s Ring is partnering with Flock—a network used by ICE, police & feds—in a major privacy red flag 🚨

Cops can now request Ring footage, expanding AI surveillance into our neighborhoods 🎥
Flock tracks vehicles & uses biased AI to ID people ⚠️

Given Ring’s privacy failures, this is deeply alarming 😠

🔗 techcrunch.com/2025/10/16/amaz

Amazon's Ring to partner with Flock, a network of AI cameras used by ICE, feds, and police | TechCrunch

Agencies that use Flock can request that Ring doorbell users share footage to help with "evidence collection and investigative work."

techcrunch.com · TechCrunch

0
0
0