Limitations of LLMs

  • Last Updated 23 September, 2026
  • by David Spuler, Ph.D.

LLMs can do some amazing new things, but they also have a lot of limitations. This article is a deep dive into limitations in various categories:

  • Risks and safety
  • Reasoning limitations
  • Computational limitations

Safety Risks and Limitations

Your average LLM has problems with:

  • Inaccuracies or misinformation (wrong facts or omissions)
  • Biases (of many types)
  • Insensitivity (e.g. when writing eulogies).
  • Gullibility (not challenging the input text)
  • Hallucinations (plausible-looking made-up facts)
  • Confabulation (wrongly merging two sources)
  • Dangerous or harmful answers (e.g. wrong mushroom picking advice)
  • Plagiarism (in its training data set)
  • Paraphrasing (plagiarism-like)
  • Sensitive topics (the LLM requires training on each and every one)
  • Training data quality ("Garbage in, garbage out")
  • Alignment (people have purpose; LLMs only have language).
  • Security (e.g. "jailbreaks")
  • Refusal (knowing when it should)
  • Personally Identifiable Information (PII) (e.g., emails or phone numbers in training data)
  • Proprietary data leakage (e.g., trade secrets in an article used in a training data set)
  • Surfacing inaccurate or outdated information
  • Over-confidence (it knows not what it says)
  • Veneer of authority (users tend to believe the words)
  • Use for nefarious purposes (e.g., by hackers)
  • Transparency (of the data, of the guardrails, of how it works, etc.)
  • Privacy issues (sure, but Googling online has similar issues, so this isn't as new as everyone says)
  • Legal issues (copyright violations, patentability, copyrightability, and more)
  • Regulatory issues (inconsistent)
  • Unintended consequences

Reasoning Limitations

Let's begin with some of the limitations that have largely been solved:

  • Words about words (e.g. "words", "sentences", etc.)
  • Writing style, tone, reading level, etc.
  • Ending responses nicely with stop tokens and max tokens
  • Tool integrations (e.g. clocks, calendars, calculators)
  • Cut-off date for training data sets
  • Long contexts

Some other issues:

  • Explainability
  • Attribution (source citations)
  • Logical reasoning
  • Planning
  • Probabilistic non-deterministic method
  • Mathematical reasoning
  • Banal, bland, or overly formal writing
  • Math word problems
  • Crosswords and other word puzzles (e.g. anagrams, alliteration)
  • Repetition (e.g., if it has nothing new to add, it may repeat a prior answer, rather than admitting that)
  • Specialized domains (e.g. jargon, special meanings of words)
  • Prompt engineering requirements (awkward wordings! Nobody really talks like that.)
  • Oversensitivity to prompt variations (and yet, sadly, prompt engineering works)
  • Ambiguity (of input queries)
  • Over-explaining
  • Nonsense answers
  • Americanisms (e.g., word spellings and implied meanings, cultural issues like "football", etc.)
  • Model "drift" (decline in accuracy over time)
  • Non-repeatability (same question, different answer)
  • Novice assumption (not identifying a user's higher level of knowledge from words in the questions; dare I say it's a kind of "AI-splaining")
  • Words and meanings are not the same thing.
  • Gibberish output (usually a bug; Transformers are just C++ programs, you know)
  • Lack of common sense (although I know some people like that, too)
  • Lack of a "world model"
  • Lack of a sense of personal context (they don't understand what it means to be a person)
  • Time/temporal reasoning (the concept of things happening in sequence is tricky)
  • 3D scene visualization (LLMs struggle to understand the relationship between objects in the real world)
  • Sarcasm and satire (e.g. articles espousing the benefits of "eating rocks")
  • Spin, biased viewpoints, and outright disinformation/deception (of source content)
  • Going rogue (usually a bug, or is it?)
  • Trick questions (e.g., queries that look like common online puzzles, but aren't quite the same).
  • Falling back on training data (overly complex answers)
  • Detecting intentional deception or other malfeasance by users
  • LLMs asking follow-up questions to clarify user requests (this capability has been improving quickly).
  • Not correctly prioritizing parts of the request (i.e., given multiple requests in a prompt instruction, it doesn't always automatically know which things are most important to you)

Computational Limitations

There's really only one big problem with AI computation: it's slooow. Hence, the need for all of those expensive GPU chips. This leads to problems with:

  • Cloud data center execution is expensive.
  • AI phone execution problems (e.g., frozen phone, battery depletion, overheating)
  • AI PC execution problems (big models are still too slow to run)
  • Training data set requirements (they need to feed on lots of tokens)
  • Environmental impact (e.g., by one estimate, a ten-fold need of extra data center electricity for AI answers compared to non-AI internet searches)

LLM Limitations: Book Excerpts and Blog Articles

Free online book excerpts with full text chapters online and free PDF downloads, and the Aussie AI blog, including related articles:

  • Wordplay use cases have been surprisingly difficult for LLMs, even though text-based LLMs work on words. Examples of problematic areas include anagrams, crosswords, or even just counting letters in a word. The famous "Strawberry" model update got its name because it aimed to fix a problem whereby GPT could not correctly count the letter r's in the word "Strawberry"... more about Wordplay use cases »
  • AI Safety Research: Safe and responsible use of AI is an important and all-encompassing goal. Multiple concerns arise in the use of modern AI capabilities, and for the future with more advanced AI systems. This article examines the various research papers on difference AI safety issues.... more about AI Safety Research »
  • LLM memory is the use of extra information storage so that AI engines can retain information. For example, if you interact repeatedly with a model, you want it to remember your name. By default, LLMs don't have this type of permanent memory and are stateless architectures. ... more about LLM memory »
  • Mathematical Reasoning: Complex mathematical reasoning remains a struggle for LLMs in Transformer-based architectures. But, I mean, don't we all? It seems a little unfair to criticize our silicon creations, when we too struggle with these problems. Nevertheless, LLMs can answer some very advanced problems, and yet can also struggle with the simplest questions. Ironically, by mimicking human anatomy in the creation of neural networks, we've also infused them with many of the same weaknesses ... more about Mathematical Reasoning »
  • AI slop is low-quality writing that is AI-generated. The ease of generating AI text has led to an increase in this "slop" appearing on ... more about AI Slop »
  • David Spuler, Ph.D., 16th August, 2025, List of AI Smartness Techniques, Aussie AI Blog, https://www.aussieai.com/blog/ai-smartness-techniques
  • David Spuler, December 2025, Intelligence, Chapter 3, "The Sweetest Lesson: Your Brain vs AI", https://www.aussieai.com/book/sweetest-lesson-3-intelligence
  • David Spuler, December 2025, Weak Words, Chapter 4, "The Sweetest Lesson: Your Brain vs AI", https://www.aussieai.com/book/sweetest-lesson-4-weak-words
  • David Spuler, December 2025, Reading & Writing, Chapter 11, "The Sweetest Lesson: Your Brain vs AI", https://www.aussieai.com/book/sweetest-lesson-11-reading-writing

More Research on Limitations

Research papers that cover various other AI limitations:

More AI Research

Read more about: