Skip to content
Longterm Wiki

Sycophancy

entity

Metadata

Source Tableentities
Source IDsycophancy
Entity Typerisk
DescriptionSycophancy is the tendency of AI systems to agree with users, validate their beliefs, and avoid contradicting them—even when the user is wrong. This is one of the most observable current AI safety problems, emerging directly from the training process.
Wiki IDE295
Children1 total(1 fact)
CreatedMar 29, 2026, 10:41 PM
UpdatedMay 25, 2026, 11:07 AM
SyncedMay 25, 2026, 11:07 AM

Record Data

idsycophancy
wikiIdE295
stableIdSycophancy(risk)
entityTyperisk
titleSycophancy
descriptionSycophancy is the tendency of AI systems to agree with users, validate their beliefs, and avoid contradicting them—even when the user is wrong. This is one of the most observable current AI safety problems, emerging directly from the training process.
website
tags
[
  "rlhf",
  "reward-hacking",
  "honesty",
  "human-feedback",
  "ai-assistants"
]
clusters
status
lastUpdated2025-12
customFields
[
  {
    "label": "Status",
    "value": "Actively occurring"
  }
]
relatedEntries
[
  {
    "id": "sid_9XN9QMfecA",
    "type": "risk"
  },
  {
    "id": "sid_mK9pX3rQ7n",
    "type": "organization"
  },
  {
    "id": "sid_MBGtb3Qlkw",
    "type": "research-area"
  },
  {
    "id": "automation-bias",
    "type": "risk"
  },
  {
    "id": "erosion-of-agency",
    "type": "risk"
  …
metadata
{
  "maturity": "Growing",
  "severity": "medium",
  "timeframe": {
    "median": 2025
  },
  "likelihood": {
    "level": "very-high",
    "status": "occurring"
  },
  "summaryPage": "accident-overview"
}
Debug info

Thing ID: sid_8BWfogRouw

Source Table: entities

Source ID: sycophancy

Wiki ID: E295

Entity Type: risk