Answer ยท confidence: partial
When did people first use AI without knowing, through Siri or spam filters?
Research question: When did people first use AI without knowing it, through search engines, spam filters, or voice assistants like Siri?
The records show the earliest example here is spam filtering: Sahami et al.'s 1998 work applied a Naive Bayesian classifier to anti-spam filtering 2. Naive Bayesian classifiers had been found surprisingly effective in studies dating back to the early 1990s, before their use in spam filtering 4. Bayesian spam filters rely on word probabilities, and automated handling of email like this often happens without users necessarily noticing 51. The evidence does not address search engines or voice assistants like Siri, and it does not say when people first used AI without knowing it 2.
Checked 2026-10-19. Only documented and supported facts are used.
Dig deeper on this questionSources
[1] supported
Spam filters use machine learning to identify and filter spam, which is one way automated systems in email handle messages without users necessarily noticing.
"More broadly machine learning can be used to identify and filter spam."Email spam
[2] supported
The document does not say when people first used AI without knowing it. It only describes machine learning for spam filtering as of 2000, citing Sahami et al.'s 1998 work as the earlier application of a Naive Bayesian classifier to anti-spam filtering.
"Sahami et al. trained a Naive Bayesian classifier (Duda & Hart, 1973; Mitchell 1997) on manually categorized legitimate and spam messages, reporting impressive precision and recall on unseen messages."An evaluation of Naive Bayesian anti-spam filtering (2000)
[3] supported
Bayesian filtering became a popular spam-filtering technique, and spammers then began using methods to weaken it.
"As Bayesian filtering has become popular as a spam-filtering technique, spammers have started using methods to weaken it."Email spam
[4] supported
Naive Bayesian classifiers had been found surprisingly effective in studies dating back to the early 1990s, well before their use in spam filtering.
"Several studies have found the Naive Bayesian classifier to be surprisingly effective (Langley et al., 1992; Domingos & Pazzani, 1996), despite the fact that its independence assumption is usually over-simplistic."An evaluation of Naive Bayesian anti-spam filtering (2000)
[5] supported
Bayesian spam filters work by relying on word probabilities, so a message with many words used only in spam is likely to be flagged as spam.
"Bayesian filters rely on word probabilities."Email spam