����ֱ��

When trying to determine how to sentence a guilty party, a judge will often look at precedent to determine an appropriate judgement. This can take time, as the judge and his or her staff pore over records and try to make a fair assessment.

But what if the technology existed to analyze hundreds, if not thousands, of similar cases quickly and build a fair judgement much faster?

Even better, what if this technology was affordable enough to be accessible in cases where hiring a lawyer was prohibitively expensive? What if you could use it when an eBay transaction goes wrong, or if you could use it even if you lived remotely and didn’t have access to a lawyer?

Enter the Intelligent Dispute Resolution System, a product of the Queen’s Law and Smith School of Business Conflict Analytics Lab. This AI-powered tool, already under development, would be capable of offering legal predictions to self-represented litigants along with negotiation support.

The “AI-Tribunal for Small Claims: Building an Intelligent Dispute Resolution System” project recently received $244,562 in funding from the Social Sciences and Humanities Research Council (SSHRC) to develop the first components of the pilot research: severance calculation predictive models, and an intelligent system for algorithmic employment negotiation.

“This is the core project of the Conflict Analytics Lab because it is touching upon many areas of our work: legal predictions, negotiation support, democratization of technology, and improving access to justice,” says Professor Samuel Dahan, Director of the Conflict Analytics Lab.

Dahan has partnered with Professor Yuri Levin, the Stephen J.R. Smith Chair of Analytics and Executive Director of Analytics and AI with Smith School of Business; Professor Xiaodan Zhu of Queen’s Electrical and Computer Engineering; and Professor Maxime Cohen of McGill’s Desautels Faculty of Management. Also assisting with the project are up to 15 students earning such degrees as Master of Laws, Master of Business Administration and Master of Management in Artificial Intelligence.

There are three components to the dispute resolution program – a legal component, a computer science component, and a data science component. The user inputs the relevant data into the system, and it returns a relevant suggestion. The system also learns as it works, meaning its suggestions will only improve with time and use.

The employment notice (“severance”) predictor, which the team has been working on the longest, is intended to help employees in situations where they have been terminated. By typing in their variables including industry, region, age of employee and length of employment, the system can suggest an appropriate severance amount. This means the terminated employee can use this system – without hiring an employment lawyer – to understand if they are being compensated fairly if they are let go by their employer.

While the system is beginning with employment law, Dahan sees the potential for application in small claims, family law,insurance, trademark disputes, and beyond. The only stipulations are that it must be a monetary award, and the award amount must be under $50,000.

“We are hoping to launch our first pilot project this fall – the employment notice predictor – and through a practicum course we are continuing to develop our data sets and explore new legal questions that could benefit from the application of AI,” says Dahan. “Over the next two years, we will be working with the Ontario Attorney General and the British Columbia Small Claims Tribunal to integrate this technology into their system.The idea will be to integrate the various tools, including the legal predictions, the dispute resolution, and the negotiation tool, into existing judicial procedures.”

In the future, Dahan hopes not only that people will use this tool but he hopes to hear from users who receive positive settlements and from companies who successfully integrate the platform into their online dispute resolution processes.

As Dahan puts it, “If they say, ‘you've helped us to sort out half of our customer service cases by making offers to unhappy customers, and our employees are much less overwhelmed than they used to be,’ I would call that a success.”

This article was first published on the .