Artificial intelligence can change contract negotiation – 11/28/2023 – Tech

“Lawyers are tired. They’re bored a lot of the time. Having something that does the boring work and leaves you free to focus on anticipating your strategies: that’s the solution.”

This is the managing director and head of human resources at the British company Luminance, Jaeger Glucina.

The company was founded in 2015 and specializes in artificial intelligence (AI) for professionals in the legal sector.

Before joining Luminance in 2017, Glucina trained as a lawyer and prosecutor in New Zealand.

“Professionals in the legal sector are clearly very well-trained people,” according to her.

“But in reality, they spend an enormous amount of their time analyzing [contratos].”

“It could take someone up to an hour to review a confidentiality agreement. And there could be hundreds of these documents [em um escritório] every day.”

With this in mind, Luminance is preparing to launch a fully automatic contract negotiation tool called Luminance Autopilot.

The company plans to begin beta testing among select customers in December and offer it more broadly in 2024.

I was invited to observe the tool in action at the company’s office in London.

On the table in front of me are two laptops.

The one on the left, for demonstration purposes, belongs to Luminance general counsel Harry Borovick.

The one on the right represents Connagh McCormick, general counsel for the company ProSapient, a real Luminance client.

A large screen covers the wall behind the laptops, showing the process of monitoring changes made by each party during contract negotiation.

The computers will use Autopilot to negotiate a confidentiality agreement that is acceptable to both parties.

Non-disclosure agreements (NDAs) serve to define the terms under which one organization will reveal its confidential information to the other party.

The demonstration begins. Borovick’s machine receives an email with the NDA attached, which opens in Microsoft Word. Autopilot quickly reviews the contract and starts making changes.

The six-year period is unacceptable and changed to three years. The legislation governing the contract changes from Russia to England and Wales.

The next risk clause establishes unlimited liability, that is, there is no ceiling that defines the maximum amount that Luminance will eventually need to pay, in case of breach of contractual terms.

“This is a hindrance to Harry’s business,” says Glucina.

“Therefore, the software proposed a liability limit of £1 million [cerca de R$ 6,2 milhões]. And he also softened the clause.”

“The other party had included a section about ‘exemption,’ meaning that they would be absolved of any legal liability in certain situations,” she continues.

“But the AI ​​knew this was bad, and so it protected Harry from this risk by removing the clause.”

What comes next is a contractual dance, with both AIs trying to improve conditions for their respective owners.

Borovick’s computer automatically returns the amended NDA via email and the contract opens on McCormick’s machine.

Your AI realizes that the possibility of exemption has been excluded and adds an indemnity clause.

Specifically, it transforms the maximum liability of £1 million into defined compensation between the parties, to be paid in the event of a breach of contract.

Borovick’s AI eliminates this clause upon receiving the updated contract and adds a section that states that his company is only responsible for direct losses incurred.

Both parties accept the fourth version of the contract. McCormick’s AI accepts all changes and sends the contract to the online contract signing service Docusign.

“At this point, we need to decide whether we really want the human being to sign,” explains Glucina.

“And this would literally be the only part of the work to be done by humans. We have a mutually agreed upon contract, completely negotiated by AI.”

The entire process took just a few minutes.

“The idea is to reduce the delays that are often seen when people fail to receive something in their inbox, or when they are too busy with another task”, explains Glucina.

What about jobs?

Autopilot is an evolution of Luminance’s copilot tool, which color-codes contract clauses so legal professionals can analyze them in Word.

Acceptable clauses are in green, unacceptable clauses in red and non-standard clauses in yellow. The tool can also rewrite clauses with AI, based on its knowledge of contracts accepted by the company in the past.

Other companies, such as Lexcheck, Genie AI and ThoughtRiver, also offer contract analysis technologies. But Luminance is the first to announce an autopilot.

Luminance’s system is based on a large language model (LLM), the same one that serves as the basis for the popular ChatGPT text generation tool.

The main difference is that Luminance’s tools were trained on more than 150 million legal documents, while ChatGPT uses public content available on the internet.

Knowledge banks created by Luminance users include their own signed documents, so the software can learn which contractual terms are typically accepted by that specific company.

After watching the demo, I spoke with ProSapient General Counsel Connagh McCormick.

Your company seeks experts for investors, consultants and other clients interested in researching specific market sectors.

In his team, three people work with contracts and there are always 20 to 30 ongoing negotiations with clients.

Some take just 48 hours, but others last up to 12 months.

ProSapient uses Luminance solutions to accelerate contract analysis. And McCormick is looking forward to testing the new Autopilot when it becomes available.

“Some people ask ‘how do you feel about AI doing all the trading?'”

“Everything ends up in the hands of the two lawyers, to sign,” he says.

“At this point, I will read the contract and the other lawyer will read the contract. If I disagree with any point, I have the opportunity to highlight the issue. I am not committed to anything that has been done by the AI.”

But what about the risk to jobs? “You’re always going to need that human step there,” he says.

“Part of the reason people go to lawyers is trust. It’s much harder to get AI to act responsibly than a person. AI means a lawyer’s time will be spent doing something more interesting, more valuable.” , he adds.

Nick Emmerson, president of the Law Society of England and Wales, agrees that lawyers will continue to be needed.

For him, “at the moment and, probably, until the distant future, AI will not be able to completely replace the legal experience offered by legally qualified professionals.”

“This is because customers have different needs and vulnerabilities that cannot yet be met by a machine, and human judgment is needed to ensure that automatic decisions do not result in potential false positives,” he adds.

“There is also a negotiation and, ultimately, bargaining technique that AI is unlikely to master.

“With any technological innovation, what it means to be a lawyer is likely to evolve, as are the types of jobs and skills required,” concludes Emmerson.

Text originally published here