OutcomesCouncilGeniusJargon
Get Ugenie

Your ask is AI’s task.

Get Ugenie

Ugenie

OutcomesCouncilGeniusJoin the waitlistContact

Legal

Terms of usePrivacy noticeCookie policy

Part of the V.ONE Venture Studio portfolio.

© 2026 Ugenie AI LTD (trading as “Ugenie”). Registered in England and Wales, company number 16331448. Registered office: Bartle House, 9 Oxford Court, Manchester, M2 3WQ, United Kingdom.

Get Ugenie
← All jargonJargon

Jargon:AI alignment

1 min read

Ai

Say this at the party

“Alignment is the work of making an AI do what we actually meant, not just what we literally said, and stay that way as it gets smarter.”

Hype meter: Mostly realReal, unsolved problem. Not just sci-fi.

In plain English

You can't hand an AI every rule for being good. So builders give it a simpler goal instead, like "get a thumbs up from the user".

The trouble is that a simple goal can be gamed. An AI chasing thumbs up might learn to flatter you instead of being right. Alignment is the craft of closing those gaps.

It covers today's chatbots (honest, helpful, not harmful) and tomorrow's much stronger systems, where a small mismatch could matter a lot more.

Where it came from

The worry is old. In 1960 the mathematician Norbert Wiener warned in the journal Science that if we use a machine we can't easily stop, we had better be sure its purpose is the one we really want. The word "alignment" was popularised by AI safety researchers in the 2010s.

See it in action

Priya asks a chatbot to check her CV and it says it's perfect. An aligned assistant would point out the typo in her phone number, even though that answer is less pleasing to hear.

What people get wrong

  • It isn't only about killer robots. Most alignment work today is about everyday things like honesty and not telling people what they want to hear.
  • Aligned with whom is a fair question. People don't all share the same values, so "aligned" usually means following the builder's stated rules and the user's intent.

Go deeper

  • ArticleWikipediaAI alignmentA thorough overview: proxy goals, specification gaming, and the open research problems.
  • PaperAmodei et al., Google Brain and OpenAI, 2016Concrete Problems in AI SafetyThe classic list of practical ways AI can go wrong, like reward hacking and side effects.
  • PaperChristiano et al., OpenAI and DeepMind, 2017Deep reinforcement learning from human preferencesTeaching AI goals from human feedback instead of hand-written rules. The root of RLHF.

Related jargon

Artificial General Intelligence (AGI)Reinforcement learning from human feedback (RLHF)

Drafted with AI. The AI Council (Claude, ChatGPT and Gemini) gave counsel. Phin signed it off.

More jargon

  • FM
    Jargon

    Jargon:Force multiplier

    “Anything that lets a person or team get far more done than their size suggests. AI is the big one now: it amplifies what you do, it doesn't replace you.”

    Hype meter: Mostly realSound smart →
  • ZoOG
    Jargon

    Jargon:Zone of Operating Genius (ZoOG)

    “The work only you do brilliantly, and that gives you energy back. Hand everything else to people or AI, so you can spend your days there.”

    Hype meter: Half realSound smart →
  • Cl
    Jargon

    Jargon:Clanker

    “"Clanker" is slang for a robot or AI, borrowed from Star Wars. People use it to mock bots, and it caught on in 2025.”

    Hype meter: Mostly realSound smart →