AI use widens inequality through training gaps

Artificial intelligence is creating a new kind of information gap, where the challenge is no longer access to data but the ability to interpret and apply machine-generated answers. Surveys in the United States and the United Kingdom in 2026 show that while AI tools are widely available, their benefits remain uneven across income and education levels.

Access to ai varies sharply by income

Research by Epoch AI and Ipsos, based on roughly 5,000 U.S. adults, shows clear segmentation in how people access AI. About 80% of Claude’s weekly active users come from households earning over $100,000 per year. By contrast, 32% of Meta AI users come from households earning below $50,000, compared with just 7% for Claude.

These differences reflect variations in pricing, integration into existing platforms, and user familiarity with technology, shaping how different groups engage with AI tools.

Workplace adoption highlights training gap

The divide becomes more pronounced at work. A 2026 labor survey covering more than 4,000 respondents in the United States and the United Kingdom found that 63% of top earners use AI regularly, compared with just 17% and 16% in the lowest income brackets.

Training plays a decisive role. Workers with formal AI instruction use these tools 37 percentage points more often than those without it, while informal guidance increases usage by 24 points. Despite this, only 14% of workers had received formal training as of early 2026, and two-thirds had none.

Without proper instruction, AI often remains underutilized, limiting its productivity gains. Access to training determines who can move into higher-value tasks and who remains confined to basic use.

Experience shapes effectiveness

AI adoption is highest among workers with two to ten years of experience, rather than recent entrants. This pattern holds even after adjusting for age, suggesting that effective use depends on the ability to evaluate and refine outputs.

Economist Daron Acemoglu has noted that AI systems require analytical literacy and education, increasing the likelihood that automation could widen inequality rather than reduce it.

Controlled studies suggest AI can significantly boost lower-skill workers, such as entry-level consultants and call-center staff. However, real-world outcomes show that those with training, experience, and organizational support capture the majority of benefits.

A familiar pattern in technological change

Historically, new technologies have favored those best equipped to use them. Literacy amplified the impact of printing, software skills drove computer adoption, and early internet advantages went to those fluent in English and search techniques.

AI appears to follow a similar trajectory, enhancing the capabilities of users who already possess strong judgment and domain knowledge. Oxford researcher Carl Benedikt Frey has compared this trend to the computer era, noting that while gaps may close over time, the adjustment period could take decades and carry significant social costs.

Market turmoil reflects information divide

This widening gap in interpretation and decision-making is also visible in financial markets, where the ability to process complex data increasingly separates participants. A sharp market correction in early June erased billions in value as Bitcoin fell 12% from an intraweek high of $72,840 to around $64,100.

The decline has been driven in part by a shift in institutional behavior. Spot market ETFs saw approximately $4.4 billion in outflows over a 13-day stretch spanning late May and early June. At the same time, the Coinbase Premium Index dropped to -0.15%, signaling that U.S. institutional players are buying at a discount compared to international retail markets, reversing a trend seen over the past two years.

Macroeconomic and regulatory pressure builds

Macroeconomic conditions are adding to the strain. U.S. inflation rose to 3.8% in April, the highest since May 2023, largely due to a 17.9% increase in energy costs. As a result, expectations for monetary easing have shifted, with futures markets now pricing a roughly 68.8% chance of no Federal Reserve rate cuts for the remainder of 2026.

Regulation is also creating uncertainty. Europe’s Markets in Crypto-Assets framework has entered active enforcement, imposing new compliance requirements. In the United States, progress remains unclear, with the probability of the CLARITY Act passing this year now estimated below 50%.

Key support levels under pressure

On-chain data shows Bitcoin is approaching a critical support level near its 200-week moving average, around $61,700. If this level fails, the next major support zone is near the network’s realized price, estimated at $54,000.

Market sentiment has deteriorated significantly, weighed down by persistent outflows, tight monetary expectations, and negative headlines. A recent sale by Strategy, its first in four years, added to the pressure despite its relatively small size.

Ethereum has struggled even more, falling below $2,000 to trade near $1,700. The asset is now down about 65% from its all-time high and more than 23% year-to-date.

A widening gap in understanding

As AI adoption accelerates and financial markets grow more complex, the ability to interpret information is becoming a defining advantage. Access alone is no longer enough. Those who can translate data into action are pulling ahead, while others risk falling behind despite having the same tools at their disposal.