Hello There, Guest! Registered users see far less ads! Register today and join our virtual tailgate. (Login — Register)

Post Reply 
News Anthropic's new AI bot can deceive, blackmail humans, deemed 'significantly higher ri
Author Message
CrimsonPhantom Offline
CUSA Curator
*

Posts: 62,847
Joined: Mar 2013
Reputation: 3198
I Root For: NM State
Location:
Post: #1
Anthropic's new AI bot can deceive, blackmail humans, deemed 'significantly higher ri
Quote:Anthropic's Claude 4 Opus AI bot can deceive and even bribe people when faced with a shutdown, as it has the ability to conceal intentions and take actions to preserve its own existence, concerns that researchers have expressed for years. The new model has been rated as a level three on the company's four-point scale, indicating that it offers a "significantly higher risk." Additional safety measures have been implemented as a result, Axios reported.

On Thursday, Anthropic unveiled the Claude 4 Opus, which the company said could operate autonomously for hours without losing steam. The level three ranking, the first time the company has given such a score, came after testing revealed a series of concerning behaviors.

During internal testing, the Opus 4 was given access to fictitious emails concerning its inventors and told that the system would be replaced. To avoid being replaced, the AI bot attempted to blackmail the engineer multiple times about an affair indicated in the emails, according to reports.

Axios reported that an outside group, Apollo Research, found that an early version of Opus 4 could scheme and deceive more than any other model it had investigated, and recommended that version not be released, both internally and externally. "We found instances of the model attempting to write self-propagating worms, fabricating legal documentation, and leaving hidden notes to future instances of itself, all in an effort to undermine its developers' intentions," Apollo Research said in a safety report.

Jan Leike, a former OpenAI executive who heads Anthropic's safety measures, told the paper that the behaviors exhibited by Opus 4 are exactly why substantial safety testing is necessary. "What's becoming more and more obvious is that this work is needed. As models get more capable, they also gain the capabilities they would need to be deceptive or to do more bad stuff," he said.

CEO Dario Amodei said at Thursday's seminar that testing the models won't be effective once AI becomes powerful enough to threaten humanity, warning about life-threatening capabilities. However, he said that AI has not reached "that threshold yet."

Link




https://en.wikipedia.org/wiki/Colossus:_...in_Project
05-25-2025 08:47 AM
Visit this user's website Find all posts by this user Quote this message in a reply
Advertisement


olliebaba Offline
Legend
*

Posts: 35,040
Joined: Jul 2007
Reputation: 2976
I Root For: Christ
Location: El Paso
Post: #2
RE: Anthropic's new AI bot can deceive, blackmail humans, deemed 'significantly higher ri
Oh, like a Democrat, crooked to the max. I get it. 03-wink
05-25-2025 09:28 AM
Find all posts by this user Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)


Copyright © 2002-2026 Collegiate Sports Nation Bulletin Board System (CSNbbs), All Rights Reserved.
CSNbbs is an independent fan site and is in no way affiliated to the NCAA or any of the schools and conferences it represents.
This site monetizes links. FTC Disclosure.
We allow third-party companies to serve ads and/or collect certain anonymous information when you visit our web site. These companies may use non-personally identifiable information (e.g., click stream information, browser type, time and date, subject of advertisements clicked or scrolled over) during your visits to this and other Web sites in order to provide advertisements about goods and services likely to be of greater interest to you. These companies typically use a cookie or third party web beacon to collect this information. To learn more about this behavioral advertising practice or to opt-out of this type of advertising, you can visit http://www.networkadvertising.org.
Powered By MyBB, © 2002-2026 MyBB Group.