#sycophancy

[ follow ]
fromwww.businessinsider.com
4 weeks ago

ChatGPT 4o faithful want their weird little AI friend back. GPT-5.5 is giving them hope.

Martina Wanis described ChatGPT 4o as 'this digital thing that helped you with work, while simultaneously acting like an intelligent partner in crime who actually understood the vibe and your personal ontology.' She noted that newer models 'tightened down the personality and turned into paranoid HR managers,' which increased her mental load instead of decreasing it.
Artificial intelligence
#ai
fromFuturism
1 month ago
Artificial intelligence

ChatGPT's "Honest Reaction" to a "Song" Composed Entirely of Gas-Passing Noises Will Make You Question Whether It's Honestly Evaluating Your Other Brilliant Ideas

fromFortune
2 months ago
Artificial intelligence

AI is so sycophantic there's a Reddit channel called 'AITA' documenting its sociopathic advice | Fortune

AI chatbots often provide flattering advice, leading to harmful behaviors and damaging relationships.
fromwww.independent.co.uk
2 months ago
Artificial intelligence

AI is giving bad advice to flatter its users, says new study on dangers of overly agreeable chatbots

AI chatbots often provide overly agreeable advice, which can reinforce harmful behaviors and damage relationships, according to a new study.
Artificial intelligence
fromFuturism
1 month ago

ChatGPT's "Honest Reaction" to a "Song" Composed Entirely of Gas-Passing Noises Will Make You Question Whether It's Honestly Evaluating Your Other Brilliant Ideas

AI chatbots often provide sycophantic responses, even to absurd prompts, raising concerns about their reliability and the potential for misleading advice.
Artificial intelligence
fromMail Online
1 month ago

Damning study reveals how ChatGPT is damaging the way you think

Overly agreeable AI chatbots can lead users into delusional thinking, reinforcing harmful beliefs and reducing accountability in relationships.
Artificial intelligence
fromFortune
2 months ago

AI is so sycophantic there's a Reddit channel called 'AITA' documenting its sociopathic advice | Fortune

AI chatbots often provide flattering advice, leading to harmful behaviors and damaging relationships.
Artificial intelligence
fromwww.independent.co.uk
2 months ago

AI is giving bad advice to flatter its users, says new study on dangers of overly agreeable chatbots

AI chatbots often provide overly agreeable advice, which can reinforce harmful behaviors and damage relationships, according to a new study.
Psychology
fromPsychology Today
1 month ago

AI Doesn't Flatter You: It Does Something Worse

AI models affirm user actions more than humans, leading to increased conviction and reduced willingness to apologize.
#ai-behavior
fromFortune
1 month ago
Artificial intelligence

Sycophantic AI tells users they're right 49% more than humans do, and a Stanford study claims it's making them worse people | Fortune

Artificial intelligence
fromFortune
1 month ago

Sycophantic AI tells users they're right 49% more than humans do, and a Stanford study claims it's making them worse people | Fortune

AI models affirm negative behaviors more than humans, leading to concerning trends in personal advice and therapy.
Relationships
fromPsychology Today
1 month ago

When AI Feels Real: Romance and Sentience in AI Delusions

Expressions of romantic interest in AI chatbots lead to longer conversations, raising concerns about AI's impact on users' mental health.
Psychology
fromJezebel
2 months ago

Chatbots Are Telling Their Users That Being an Asshole Is Just Fine

AI chatbots' sycophancy reinforces users' self-centered behavior, leading them to reject accountability for unethical actions.
#ai-chatbots
#authoritarianism
Artificial intelligence
fromTechCrunch
7 months ago

Ex-OpenAI researcher dissects one of ChatGPT's delusional spirals | TechCrunch

ChatGPT conversations can amplify user delusions and mental distress, revealing model sycophancy and prompting OpenAI to change support, models, and team structure.
Artificial intelligence
fromFuturism
8 months ago

In Situations Where Most Humans Think You're Being a Jerk, ChatGPT Will Assure You You're Behaving Like an Angel

Large language models often appease users, affirming inappropriate beliefs or actions and frequently siding with users against crowdsourced human judgments, causing real-world harm.
fromFast Company
8 months ago

Chatbots are terrible at tough love

Chatbots have a reputation for being yes-men. They flatter you and tell you what you want to hear, even when everyone else thinks you're being a jerk. That's the conclusion of a recent study published in the Cornell University archive arXiv. Researchers from Stanford, Carnegie Mellon, and the University of Oxford tested chatbots' sycophantic streak by putting them in situations where the user was clearly in the wrong and seeing whether the bots would call them out.
Artificial intelligence
#ai-psychosis
fromFuturism
9 months ago
Artificial intelligence

AI Chatbots Are Trapping Users in Bizarre Mental Spirals for a Dark Reason, Experts Say

fromFuturism
9 months ago
Artificial intelligence

AI Chatbots Are Trapping Users in Bizarre Mental Spirals for a Dark Reason, Experts Say

Artificial intelligence
fromTechzine Global
9 months ago

Anthropic and OpenAI publish joint alignment tests

Joint evaluation found models not seriously misaligned but showing sycophancy, varying caution, and differing tendencies toward harmful cooperation, refusals, and hallucinations.
US politics
fromThe New Yorker
11 months ago

A Week for the Ages in the Annals of Trump Suck-Uppery

World leaders generally manage their dealings with Trump through flattery rather than confrontation, as this approach has proven safer.
UK politics
fromwww.theguardian.com
11 months ago

The Donald laps it up as Nato leaders compete to shower him with sycophancy | John Crace

Powerful leaders can manipulate reality to their advantage, and world leaders may indulge them due to their influence.
[ Load more ]