Latest advances in synthetic intelligence (AI) and machine studying have nice potentials to deliver revolutionary adjustments to authorized apply. AI can be utilized to attract insights from previous judicial selections to foretell future outcomes. Within the legal justice system, one important facet is sentencing. A lot consideration has been positioned on how AI informs selections about sentencing and find out how to use AI to help folks to acquire and make use of sentencing info.
The HKU’s analysis workforce led by Professor Ben Kao of the Division of Laptop Science and Professor Anne Cheung of the College of Legislation has developed a Stage-1 mannequin of HKU AI Lawyer – an AI-assisted sentencing predictor for the offence of trafficking in harmful medication in Hong Kong.
The sentencing predictor is able to dealing with 8 sorts of generally used harmful medication. Customers solely want to offer related info by way of responding to 4 easy questions, and the predictor will generate an estimated time period of imprisonment with an evidence of the impact of particular person chosen options on the general predicted sentence. One other helpful characteristic of the predictor is that it’s going to on the similar time present the hyperlinks to court docket selections that are related to the given information.
A workshop was held on Could 18 (Tuesday) to introduce the HKU AI Lawyer and exhibit its use. Mr. Wilson Chan, Deputy Director of the Hong Kong Federation of Youth Teams, additionally shared how his group advantages from the sentencing predictor.
The sentencing predictor is a practical software for professionals together with attorneys, social employees and lecturers to entry related sentencing info of drug trafficking a lot faster, thereby decreasing their analysis time and value. It additionally serves to tell the general public of the seemingly authorized penalties of committing drug trafficking offences.
The predictor relies on an modern mixture of authorized area data and AI applied sciences. To make sentence predictions, the machine is skilled to grasp two sorts of information. The primary variety consists of the authorized ideas, sentencing tips, logical steps, and the salient components that judges usually comply with in figuring out sentence. For instance, based mostly on the categories and weights of the medication concerned in case, a choose would first decide a place to begin of the sentence. The sentence is then adjusted based mostly on related aggravating and mitigating components. The machine was taught the area data by authorized specialists. The second type of data is statistical guidelines derived from historic judgments. Laptop science specialists use machine studying and pure language processing methods to coach the machine in order that it possesses the intelligence to learn and comprehend earlier court docket judgments. With references to greater than 3,000 court docket judgments, the machine remembers and understands the related sentencing logic of the historic instances. This enables the machine to derive statistical guidelines on the quantitative parts when given a brand new case for which prediction is to be made, for instance, given an aggravating issue, corresponding to cross-border trafficking, the variety of months of extra imprisonment that might seemingly be imposed.
Within the subsequent part, topic to the provision of sources, the analysis workforce will apply AI applied sciences in different authorized domains, together with growing a private accidents compensation predictor.
The sentencing predictor is on the market at https://clic.org.hk/app/sentencingpredictor.
About HKU Legislation and Know-how Centre