Is interest in speech recognition growing?

Interest in the general topic is not growing. Interest in specific AI assistants is. The English Wikipedia article "Speech recognition" had 70,725 page views in the 12 months to September 2026 ¹. Its monthly count was 18,883 in July 2015 and 5,624 in September 2026. The readers have not left voice technology. They now look up the product they use, not the field behind it.

The data behind this page covers 39 data points ², each a yearly or monthly view count for one article about speech recognition, dictation, voice assistants or typing.

The classic topic article lost readers since 2015

The monthly count for "Speech recognition" fell from 18,883 in July 2015 to 5,624 in September 2026. The German, French, Japanese, Spanish, Italian and Portuguese articles on the same subject show the same direction. The Japanese article, "音声認識", went from 2,724 in July 2015 to 300 in September 2026. The French one, "Reconnaissance automatique de la parole", went from 912 to 266.

Two neighbors are flatter. "Speech synthesis" rose from 10,170 in July 2015 to 12,941 in September 2026, and it had 147,894 views in the latest 12 months ³. "Typing" barely moved, from 5,947 to 5,564 over the same dates.

Siri is still read, and recent growth is in AI assistants

The "Siri" article had 307,597 views in the latest 12 months ⁴. Its monthly count fell from 82,197 in July 2015 to 40,122 in September 2026, but September 2026 was higher than the same month a year earlier, so it is rising from a low point.

The steepest climbs belong to newer products. "Google Gemini" went from 595 views in October 2023 to 89,234 in September 2026, with 1,093,862 views in the latest 12 months ⁵. "Microsoft Copilot" went from 41 in March 2023 to 38,318, and 644,119 over the year ⁶. "Amazon Alexa" went from 59 in December 2015 to 15,740. The speech model "Whisper" went from 664 in August 2023 to 4,021.

Not every new name keeps growing. "Grok (chatbot)" had 1,037,826 views over the year ⁷, but its September 2026 count was below the same month a year earlier.

The most read topics are chatbots, not dictation

The table lists the topics with the most views in the latest 12 months.

Showing 8 of 39 rows. See every row in the dataset.

Typing topics that once drew many readers have shrunk. "Carpal tunnel syndrome" went from 110,650 in July 2015 to 13,173 in September 2026. "Words per minute" went from 22,282 to 5,280. "Repetitive strain injury" went from 23,187 to 3,983.

What these numbers do and do not measure

Page views count visits, not unique readers. They include people who reached an article by accident. A news event or a search engine change can move a count without any change in real interest. Search engines and AI answers now often show the facts directly, so fewer people may click through to an encyclopedia page even if they care as much as before. A falling count can therefore mean "answered elsewhere", not "lost interest".

The data counts only views by people, and it excludes known crawlers and automated agents. It covers Wikipedia only, in English and a few other languages. It says nothing about how many people dictate, and it does not measure usage of any product. Each article also has a fixed scope: "Voice user interface" and "Speech-to-text" are small articles, so their low counts partly reflect where people land, not how many care about the idea.

What it means if you are choosing how to dictate

Falling interest in the topic does not mean the technology got worse or that people stopped using it. The attention moved to products, and the products are built on the same recognition research. For a short history of how the field got here, read the history of speech recognition.

If you are choosing a tool, do not use Wikipedia readership as a proxy for quality. A high count shows a product is in the news. Compare accuracy, speed, privacy and price instead, and see our market comparison of dictation tools. Start with the free tools built into your phone or computer, and move to a dedicated app only if they fall short.

Where to read more

• The full dataset lists every article, its last month, the same month a year earlier and its history.
• Speech recognition overview explains how the technology works and where it is used.
• Is Dragon NaturallySpeaking still popular? looks at one older product in the same data.