AI Will Change Legal Services. It Will Not Remove the Need for Lawyers.

The most common question asked about legal technology is also the least useful one: will AI replace lawyers? It is unhelpful because it treats legal work as a single task rather than a profession made up of very different activities — some of which software handles well, and some of which require a person who can be held responsible for the outcome.

A more accurate question is narrower. Which parts of the journey to legal help are slow, repetitive and structural, and which parts require judgment, context and accountability? The first group is where AI is already useful. The second is where qualified lawyers remain irreplaceable.

AI is already part of legal work

Adoption is no longer speculative. More than 80% of organisations surveyed by Thomson Reuters reported using generative AI weekly. In practice, that use clusters around intake, summarisation, first-pass research and document review — work that benefits from speed and consistency.

None of those tasks is the practice of law in itself. They are the preparation around it: organising facts, finding relevant material, and removing the manual work that sits between a person’s problem and a professional’s attention.

Where the profession draws the line

Practitioners themselves are considerably more cautious about advice than about assistance. Only 17% of surveyed legal professionals were ethically comfortable allowing AI to give legal advice. That reluctance is not technophobia; it reflects how legal responsibility works.

Professional responsibility cannot be delegated to software

When an attorney advises a client, they accept duties: competence, confidentiality, loyalty, candour and the obligation to act in the client’s interest. Those duties attach to a person and a licence. They can be enforced, insured and, where breached, remedied.

The American Bar Association confirms that lawyers using AI retain their professional duties. A tool does not absorb responsibility; the practitioner remains accountable for the work product, the advice given and the confidentiality of client information.

This is the structural reason AI does not replace lawyers. An AI system cannot be admitted, cannot owe a fiduciary duty, cannot be sanctioned, and cannot be sued for negligence. Where responsibility is required, a responsible person is required.

AI can help people reach the door. A qualified lawyer still needs to walk through it with them.

Legal work requires context, empathy and judgment

Most legal matters are not purely legal. A dismissal is also a household income problem. A divorce is also a parenting arrangement. A commercial dispute is also a relationship the client may need to preserve. Deciding what to pursue, what to settle and what to leave alone requires understanding what the client actually wants from the outcome.

Judgment of that kind is not pattern completion. It involves weighing risk against cost, reading the other side, knowing how a particular forum tends to behave, and telling a client something they do not want to hear.

AI output requires human oversight

Generative systems produce fluent text regardless of accuracy. They can misstate authority, blur jurisdictional differences, or present an outdated position with complete confidence. In a legal context, a plausible error is more dangerous than an obvious one, because it is harder to detect and easier to rely on.

Oversight is therefore not an optional safeguard. Every output that affects a person’s rights should pass through a professional who can verify it against the applicable law and the facts of that specific matter.

The public wants both: easier access and human accountability

Consumer research points in the same direction. Legal Services Board research found that 70% expected AI to make legal services easier to use, 66% expected better accessibility and 64% expected greater affordability.

The same research shows those expectations come with conditions. Consumers still expected human oversight, informed consent, accuracy, privacy and access to redress. In other words, the public is open to technology improving access, and unwilling to give up the protections that come with a regulated professional.

How Attora uses AI

Attora is not being built to replace attorneys. It is being built to help people find them.

Our AI is designed to do the structural work: listen to what happened in the person’s own words, organise it into a clear brief, and identify the practice area, jurisdiction, location and experience the matter calls for. It does not diagnose claims, predict outcomes or give legal advice. It prepares information so that a qualified professional can assess a matter quickly and decide whether it fits their practice.

For attorneys, that changes the economics of intake. Attention is the profession’s scarcest resource, and spending it on well-described matters inside the right practice area serves practitioners and the public simultaneously.

No matter how capable AI becomes, people will continue to need trusted professionals who understand context, exercise judgment and take responsibility. Attora is building the platform that helps people find those professionals.

This article is general commentary and does not constitute legal advice. For advice on a specific matter, consult a qualified legal practitioner.