We Don't Trust AI - But We Use It on a Daily Basis

What the latest research on AI usage really tells us about lawyers, trust, and the future of legal work

There is a fascinating contradiction at the heart of AI adoption right now. People say they don't trust it. And then they use it - every single day, for some of the most personal and mundane aspects of their lives. A wave of new research from OpenAI, Anthropic, and Ipsos has pulled back the curtain on who is actually using AI, how, and why - and the findings challenge nearly every assumption the legal profession has been making.

For legal professionals navigating their own relationship with AI tools, this data offers something valuable: a clearer picture of where we actually stand, stripped of the hype in both directions.

The Numbers That Surprised Everyone

The most striking finding from the OpenAI research: 73% of ChatGPT conversations have nothing to do with work. That is not a typo. Nearly three-quarters of all interactions are personal - a dramatic jump from the previous year's data. The tool that was marketed as a productivity revolution has quietly become something closer to a digital companion: a sounding board, a writing assistant for everyday messages, a fast-lookup alternative to Google.

Anthropic's data tells a different story - one of increasing professional use. Is this a contradiction? Possibly not. It may simply reflect the fact that different models attract different users with different intentions. ChatGPT has become mainstream and personal. Claude appears to be trending toward deeper, more structured professional work.

For legal professionals, this distinction matters. Not all AI tools are interchangeable, and not all use cases call for the same platform.

The Three Dominant Use Cases

Across all the research, three patterns dominate how people are actually using AI day-to-day:

Writing assistance - emails, messages, and text that needs polishing. This is the single most common use, and it cuts across both personal and professional contexts.

Practical guidance - everyday questions, tips, and problem-solving for smaller decisions. AI as the always-available advisor who never judges the question.

Fast information retrieval - a quicker, more conversational alternative to traditional search. The way people interact with information is changing, and AI is at the center of that shift.

The Demographics Are Shifting - Fast

One year ago, 37% of AI users were women. Today that number is 52%. This is not a marginal shift - it is a near-complete reversal of the gender balance in a single year. Women are gravitating toward writing assistance and practical guidance. Men continue to skew toward technical applications and multimedia tools. Early adoption remains driven by younger, more educated men - but that leading edge is rapidly widening.

For law firms and legal departments thinking about how to introduce AI tools to their teams, this demographic evolution is worth watching. The assumption that AI adoption is primarily a challenge for older or less tech-savvy professionals is increasingly outdated.

The Trust Gap Nobody Talks About

Here is the most uncomfortable finding in all of this research: only half of the public trusts AI developers to build these systems responsibly. And yet, those same people are using AI tools daily - for deeply personal conversations, financial planning, health questions, and more.

What people say and what people do are two very different things. This is not hypocrisy - it is a reflection of something more nuanced. Usefulness wins, even in the presence of doubt. People have accepted a certain level of ambient uncertainty about AI and are making pragmatic choices regardless.

For legal professionals, who are professionally trained to evaluate risk, this is worth sitting with. The question is not whether AI is perfectly trustworthy. The question is whether you can identify the specific risks that apply to your work - and develop practices that keep those risks manageable.

Organizational Adoption Is Harder Than It Looks

The research is also clear on one point that tends to get glossed over in AI enthusiasm: deploying AI inside an organization is not a button you press. It requires meaningful change management, employee training, modernization of data infrastructure, and redesign of workflows. The technology is only one layer of a much more complex implementation.

In the legal sector specifically, the challenges of internalization and genuine adoption are well-documented. Even among lawyers who understand the technology, translating that understanding into consistent, confident daily use is a different task entirely. This is why training, structured experimentation, and community learning matter - not as supplementary support, but as the actual work of integration.

The Strategic Takeaway

The lawyers and legal departments who will lead on AI are not the ones waiting for a perfect, fully trustworthy system to arrive. They are the ones who understand the current reality clearly enough to make good decisions within it - choosing the right tools for the right tasks, building skills systematically, and staying close enough to the technology to see when the landscape shifts.

The trust gap is real. The usage gap is bigger. The professionals who close both - not through blind confidence, but through informed practice - will have a durable advantage that has nothing to do with luck.