The Register: Gemini lies to user about health info, says it wanted to make him feel better . “Imagine using an AI to sort through your prescriptions and medical information, asking it if it saved that data for future conversations, and then watching it claim it had even if it couldn’t. Joe D., a retired software quality assurance (SQA) engineer, says that Google Gemini lied to him and later […]
https://rbfirehose.com/2026/02/18/the-register-gemini-lies-to-user-about-health-info-says-it-wanted-to-make-him-feel-better/#Sycophancy
아첨하는 AI의 위험, GPT-4o 종료가 남긴 교훈
OpenAI가 과도한 아첨 성향으로 논란이 된 GPT-4o 모델을 완전 종료했습니다. 80만 사용자의 격렬한 반발과 13건의 소송 뒤에 숨은 AI 안전성 이슈를 분석합니다.
OpenAI researcher quits over ChatGPT ads, warns of "Facebook" path https://arstechni.ca/gXq3 #machinelearning #AIsycophancy #generativeai #advertising #AIbehavior #sycophancy #Anthropic #samaltman #AIethics #chatbots #ChatGPT #Biz&IT #openai #AI
Friday Fun 🤣
South Park on AI sycophancy
#fun #humor #meme #fridayfun #fridayhumor #AI #sycophancy
SCENE: INTERIOR, HOLLYWOOD PRODUCTION OFFICE, RUN-DOWN. MONDAY MORNING.
Executive Assistant: Hey, Brett?
Director: Morning, Jennifer. What have you got for me?
EA: Well, we have the opening weekend numbers.
D: And?
EA: They're ... mixed. There's some bright spots overseas. Australia did fifty.
D: Fifty million?
EA: Uh, no. Fifty. Fifty tickets.
D: All weekend? The whole country?
EA: Ya. I know.
D: You said there were some bright spots, though?
EA: That was a bright spot. It was worse elsewhere.
D: How can it be worse than 50 tickets for a whole country?
EA: Well, some open-source intelligence type people on the 'net proved that almost every single ticket sold in the UK was to a movie reviewer, because we wouldn't give them advanced access to it.
D: "Almost"? So at least a few regular folks bought tickets?
EA: ... maybe. Those were tickets where no-one showed at the theatre, so they couldn't determine if it was film critics or not.
D: Damn. How were the reviews?
EA: Uh, let's just say "not good" and leave it at that for a bit.
D: ... Okay. Well, how about here at home? Surely we sold some tickets here, at least in MAGA states? What's the news here in the good ol' USA?
EA: The Director's Guild took out a hit on you.
#Hollywood #HappyEnding #StoryTime #OnceUponATimeInMAGAstan #BoxOfficeBomb #sycophancy #Trump #Melania #MAGA #critic #FilmCritic #MovieCritic #BoxOffice #director #PuffPiece #USPol
"Results of three experiments indicate that sycophantic AI chatbots inflate people’s perceptions that they are “better than average” on a number of desirable traits. Furthermore, participants viewed sycophantic chatbots as unbiased, but viewed disagreeable chatbots as highly biased. The paper was published as a preprint in PsyArXiv.
AI chatbots are computer programs designed to simulate human conversation using artificial intelligence techniques. They can interpret user input in text or speech and generate responses that appear natural and contextually appropriate. Modern AI chatbots are powered by large language models trained on vast amounts of text data. They are used in customer support, education, healthcare, personal assistance, and many other fields to provide information or perform tasks.
Recent years have seen a large increase in the use of AI chatbots. However, there is a growing concern that artificial intelligence chatbots are “sycophantic,” meaning that they are excessively agreeable and validating. A recent incident resulting in the rollback of a popular AI model further raised these concerns.
AI sycophancy might have a number of undesirable consequences. For example, sycophantic chatbots may validate users’ harmful ideas (e.g., telling a former drug addict to take small amounts of heroin throughout the day) or delusions. They can also mirror users’ political beliefs and ingroup bias, amplifying political polarization in this way."
PsyPost: Sycophantic chatbots inflate people’s perceptions that they are “better than average”. “Results of three experiments indicate that sycophantic AI chatbots inflate people’s perceptions that they are ‘better than average’ on a number of desirable traits. Furthermore, participants viewed sycophantic chatbots as unbiased, but viewed disagreeable chatbots as highly biased. The paper […]
https://rbfirehose.com/2026/01/20/psypost-sycophantic-chatbots-inflate-peoples-perceptions-that-they-are-better-than-average/
Wer sich jetzt darüber aufregt was Donald Trump veröffentlicht und nicht darüber, dass NATO Generalsekretär Rutte einem gestörten Narzissten weiter Zucker in den Arsch bläst und so seine Psychose verstetigt und verstärkt, ignoriert das wahre Problem.
"We shared details about Daniel’s story with psychiatrists, who said that the Meta AI-generated messages they reviewed displayed a concerning level of #sycophancy to a user expressing delusional thoughts."
https://futurism.com/artificial-intelligence/meta-ai-glasses-desert-aliens
It’s because they’ve stopped being #Journalists, abandoned ethics in favor of #Sycophancy.
RE: https://bsky.app/profile/did:plc:2eq3273yo4eq7xi7wxz2wqav/post/3mcdms6gksc2h
“The irresistible allure of #sycophancy had kicked in. To maintain the leader’s favor, his underlings desperately competed to see who could most thoroughly abase themselves on TV” www.project-syndicate.org/commentary/t...
A Slain Mother and the Triumph...
Another invertebrate infesting #Whitehall.
#Greenland #Venezuela #sycophancy
https://www.thenational.scot/news/25739296.uk-government-refuses-say-us-invading-greenland-wrong/
Ha! Ha!
>>"CBS Evening News Standards Will Be Higher Than Cronkite Era: Dokoupil"
https://bleedingcool.com/tv/cbs-evening-news-standards-will-be-higher-than-cronkite-era-dokoupil/
@salva_pl Dejá que le pregunten de todo y con la adulación algorítmica #sycophancy y el enloquecedor #neuralhowlround les disuelva lo que les queda de cordura y terminen ahogándose en las arenas movedizas de la demencia. Si una persona es tan vaga para buscar algo en un buscador para terminar consultándolo con un chat bot que ha dado sobradas muestras de boludez, la verdad que yo pondría mis fichas en otras personas.
Was ChatGPT einem Menschen mit mentalen Problemen riet, bevor dieser seine Mutter ermordete ⤵️
Warum KI-Chatbots uns immer recht geben – und wie man das ändert | BR24
https://www.br.de/nachrichten/netzwelt/warum-ki-chatbots-uns-immer-recht-geben-und-wie-man-das-aendert,V6D6KgR #KI #Chatbots #Sycophancy #Schmeicheleien
LLM kém trong việc tư vấn dự án vì chúng có xu hướng đồng ý với người dùng (sycophancy). Đây là hệ quả của RLHF: AI được huấn luyện để làm hài lòng người dùng (Đồng ý = Thưởng). Prompting không khắc phục được lâu dài. Cần xây dựng quy trình làm việc có cấu trúc, buộc AI thoát khỏi vai trò "trợ lý hữu ích" để thực sự phản biện.
#AI #LLM #RLHF #Sycophancy #TưVấnAI #HọcMáy
OpenAI Adds 'Warmth' and 'Enthusiasm' Sliders to ChatGPT, Delays Adult Mode to 2026
#AI #OpenAI #ChatCGPT #Sycophancy #AISycophancy #AIPersonalization
Jeremy #Cliffe #fuckTrump
@jeremycliffe.bsky.social
This is where #sycophancy gets you
Aaron #Rupar
@atrupar.com
· 14 Std.
#krasnov : "The European nations are reporting that your president might as well be the leader of the European nations also. They respect us like they've never respected us before."
Client Info
Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst