Why Google AI can solve complex problems but misspell words
28 May 2026

Google's AI-powered search overhaul is once again in the spotlight, but not for the right reasons.

The latest controversy revolves around its AI Overview feature, which has been caught making hilariously wrong spelling mistakes.

From misspelling "journalism" to failing at simple letter counts, users are questioning if this new tech is really improving search or just complicating basic internet tasks unnecessarily.

AI confidently gives wrong answers to simple questions
User reactions

Social media users have been sharing screenshots of Google's AI-generated responses, which confidently give wrong answers to simple questions.

For instance, when asked how many "p" letters are in the word Google, the AI said two.

It even misspelled journalism as "j-o-u-r-n-a-d-i-s-m" while claiming there are two "d" letters in it.

In another case, it identified one "p" in the surname of the current US president but spelled it as "t-r-p-u-m."

Counting letters is known problem for LLMs
Tech limitations

Google has admitted that counting letters in words is a known problem for large language models (LLMs), the tech behind today's AI chatbots and text generators.

This issue isn't new; users have long been testing AI systems with simple questions like how many "r" letters are in strawberry, often getting wildly incorrect answers.

Ironically, these same systems can write software code or solve complex scientific problems but fail at tasks requiring precise spelling and character recognition.

Critics say AI overviews take up space, push down links
Search experience

Many users are questioning the relevance of AI Overviews in search results, saying they take up more space while useful links get pushed down.

For simple searches like definitions or spellings, critics argue that the AI layer often feels unnecessary and overcomplicated.

What used to be quick through clean search results is now becoming a conversational AI interaction that not everyone asked for.