AI literacy could become the next big digital divide

Americans with more money and education are more likely to recognize and use artificial intelligence (AI) than those with fewer economic advantages, a new study finds.

This discovery turns ordinary tools like job screens, recommendations, and spam filters into subtle advantages for those who know how the system works.

Studying the inequality gap in AI

Internal responses collected by the Pew Research Center through the American Trends Panel, a national survey project of American adults, show that the divide is manifest in everyday digital life.

By tracking 10,087 available responses, Dr. Sai Wang linked education and income to whether people were aware of the AI around them.

Professor Wang’s team at Hong Kong Baptist University (HKBU) found that high-income and highly educated adults also use AI more and feel more familiar with it.

These patterns show that the divide wasn’t just about devices. Visibility itself shapes who gains from their digital lives.

AI often goes unnoticed

Researchers measured AI awareness using six everyday examples, including chatbots, shopping recommendations, and spam filters.

Each correct answer raised the score from 0 to 6, giving researchers an easy way to compare perceptions between groups.

Across the sample, many people were still unable to recognize some common AI-powered tools in their daily digital lives.

Even modest recognition scores reflect how hidden systems shape choices before the user is aware of the machine’s involvement.

The knowledge gap widens

The education sector stood out because it tracks AI use more strongly than income in its analysis. The team used socio-economic status to compare social advantages among respondents.

Formal schooling can develop reading, searching, and problem-solving habits that help you test tools and discover automated features.

Money remained important, but educational patterns warned that access alone could not close the gaps rooted in confidence and technological understanding.

Familiarity trumps usage

One study questioned the assumption that using an AI tool is a stronger predictor of awareness than simply feeling informed about the AI.

News, conversations at work, family conversations, exposure at school, etc. can make hidden systems easier to recognize later.

That means someone could learn how to spot an AI before opening a chatbot or purchasing a premium tool.

In everyday apps, AI is often hidden behind features that feel mundane, useful, or even neutral.

For example, some tools predict what users will want next by classifying songs, movies, posts, and products based on past behavior.

This feature simply appears as a suggestion, so many users never see the software make a selection.

Early research on algorithmic recognition helps explain why invisibility is important because people can’t challenge sorting systems that they don’t notice.

Deepfakes raise concerns

Awareness will also change the way people judge risk, especially when AI-generated images, audio, and text spread rapidly online.

Deepfakes can fool viewers because AI copies faces, voices, and text patterns. Privacy risks increase when people are unaware that tools are collecting data to predict behavior or personalize decisions.

Increased awareness doesn’t make users safer, but it does give them more opportunities to ask harder questions.

AI in job search

Jobs crystallize that gap because it can influence who job screening software focuses on.

Applicants who know that screening software may scan their resumes can match keywords more carefully and avoid unreadable formatting.

Someone else might submit the same experience in a format that never reaches a human reviewer.

This difference shows how consciousness can be a practical force without looking like a traditional technological skill.

AI literacy in schools

Strong AI literacy should start with awareness, not coding lessons or expensive tools. Community workshops can use familiar examples such as playlists, shopping suggestions, spam folders, and customer service chats.

Schools can add foundational concepts early on, so students learn that AI makes predictions from data, not magic.

“Bridging the AI awareness gap is essential, because if only those with higher incomes or higher education know about AI and how to use it, it can further deepen social inequalities,” Wang said.

Be careful with the results

Important limitations prevent our results from becoming blanket claims about all countries or all forms of AI.

This data was obtained from US adults in December 2022, years before many people had experience with new chatbots.

Cross-sectional data cannot prove which factor caused another. Still, this pattern provides educators and policy makers with a clear venue to test practical responses.