Backed by OpenAI, Harvey a co-pilot for lawyers is here to help in legal procedures.

With $5 million in funding from the OpenAI Startup Fund, the tranche through which OpenAI and its partners are investing in early-stage AI businesses tackling significant challenges, Harvey, a startup developing what it characterizes as a "copilot for lawyers," today came out of stealth. Along with other angel investors, Jeff Dean, the head of Google AI, and Elad Gil, the co-founder of Mixer Labs, took part in the round. Harvey was established by Gabriel Pereyra, a former research scientist at DeepMind, Google Brain, and Meta AI, and Winston Weinberg, a former security trust attorney at the law firm O'Melveny & Myers. Weinberg and Pereyra share a residence; when Pereyra demonstrated OpenAI's GPT-3 text-generating system to Weinberg, the latter recognized its potential for enhancing legal operations.

Harvey enables lawyers to define the task they wish to complete in simple instructions and receive the resulting outcome, as opposed to manually revising legal documents or conducting legal research. Harvey uses huge language models to do this by both understanding the users' intent and producing the right result. For example, Harvey can respond to queries such as, "Explain the distinction between an employee and an independent contractor in the Fourth Circuit." In contrast to employing a variety of intricate and specialized tools for specialized activities, Harvey offers a single and simple interface for all legal workflows.

Pereyra says in an e-mail interview, "Our platform gives lawyers a natural language interface for the current legal operations. At first glance, Harvey nearly seems like he could take the position of attorney making legal justifications and submitting drafts on demand. Pereyra, though, is adamant that this is untrue. As a natural language interface to the law, he added, "We want Harvey to act as a bridge between tech and lawyers." "Harvey will increase lawyers' productivity, enabling them to do better work and devote more time to the high-value aspects of their jobs.

In theory, it's strong stuff. It's risky, though. Given the intensely private nature of the majority of legal disputes, attorneys and law firms could be hesitant to grant Harvey access to any case files. Additionally, language models tend to use poisonous language and makeup information that would be highly offensive in court, if not downright perjurious. Harvey, which is still in beta, has a disclaimer attached to it explaining that it shouldn't be used by nonlawyers and should only be used in conjunction with qualified legal counsel.

Pereyra claims that Harvey goes to great lengths to satisfy clients' compliance concerns by anonymizing user data and destroying data after a predetermined period. He asserts that users can request the deletion of their data at any time and find solace in the knowledge that Harvey doesn't "cross-contaminate" data between customers. There is some opposition to it. To locate legal cases, help with general legal research chores, and with the brief drafting process, Casetext leverages AI, especially GPT-3. AI-powered surgical tools like Klarity remove the tediousness of contract review. To assist tenants in asserting their rights, startup Augrented once considered using GPT-3 to summarize legal notices or other sources in plain English.