[ad_1]
Source: Business Times
Article Date: 22 Nov 2023
Author: Renald Yeo
The free basic version of ChatGPT can be used to draft legal documents at no cost; the quality of the AI output will depend on the user’s own expertise and familiarity with the subject matter. However, the danger comes when contractual disputes arise.
Some small-business owners in Singapore are using generative artificial intelligence (AI) tools such as ChatGPT to draft legal documents, but lawyers warn that the risks in doing so could outweigh any benefits.
Documents such as employment contracts and supply agreements could cost thousands of dollars if drafted in full by a lawyer. In contrast, the free basic version of ChatGPT can be used to draft these documents at no cost.
Yet, the quality of the AI output will depend to a large extent on the user’s own expertise and familiarity with the subject matter. The user will have to consider if he is comfortable with crafting the prompts and, if so, to do so carefully, said Andrea Chee, director at AEI Legal. “A key concern will be not knowing what one doesn’t know.”
Existing templates for simple contracts are already available for free online, noted TSMP joint managing partner Stefanie Yuen Thio.
For complex or nuanced work, however, a lawyer should still be roped in. Said Yuen Thio: “I cannot tell you how many clients we have had to help who tried to save money by drafting their own contract, and then ran into difficulties when they found the contract was incomplete, internally inconsistent or unclear.”
The danger comes when contractual disputes arise. That is when the company would have to bear responsibility for the contract’s text, said Singapore University of Social Sciences (SUSS) law lecturer Ben Chester Cheong.
“The company directors would have to take responsibility for mismanagement or legal consequences if the poor drafting of contracts by generative AI exposes the company to civil or criminal liability,” said Cheong, who also practises at RHTLaw Asia.
“It is not a defence for company directors to say that the AI drafted the legal document, so they do not know what is going on and hope to get out of the tangle,” he added.
Automated fine print
One business owner told The Business Times that his company saved about S$3,000 by using AI to draft a supply agreement contract this year.
The business owner, who declined to be named, fed a similar document – drafted by the company’s usual law firm – into the basic version of ChatGPT.
Using that as a template, the AI was prompted to add sections relevant to the new agreement. The full AI-drafted document was then sent to the company’s law firm for final checks. The idea was to reduce the amount of billable hours, and thus save costs.
When drafting the first cut with ChatGPT, the business owner made sure to use placeholder names within the document to maintain confidentiality. The identities of the entities involved were later added manually.
Lawyers said that even with such precautionary measures, however, confidentiality concerns remain. That is because AI tools typically use algorithms that learn from their interactions with users.
“Companies should note that the generative AI bot may be mining the information these companies input into the application, and consider if confidential or sensitive information is being ‘shared’,” said Yuen Thio.
Ronald Wong, director at Covenant Chambers, added: “Parties also risk inputting their confidential information and transaction details into a system they have no control over, which could end up popping up in someone else’s output.”
For professional use only
Lawyers told BT they have not seen any impact on revenue from clients drafting their own contracts, as the solutions they offer go beyond such drafting.
Nevertheless, they noted that the use of generative AI tools in the legal industry is an “inevitable” prospect.
Covenant Chambers, for instance, is piloting the use of AI within its in-house data system for evidential and document review and analysis.
Generative AI tools developed specifically for the legal industry – as opposed to more generic ones – can be valuable, said Rajesh Sreenivasan, partner and head of Rajah & Tann’s technology, media and telecommunications practice.
He cited American startup Harvey, which builds AI tools specifically for the legal sector. “(The AI models) are trained on extensive legal document databases, making them proficient in generating initial drafts of straightforward legal documents,” he noted.
“However, it’s crucial to emphasise that even with the assistance of legal-specific generative AI systems, a legally trained professional should review the final work product,” he added.
In September, Harvey signed a two-year memorandum of understanding with the Singapore Courts to develop a generative AI program for users of the Small Claims Tribunals.
“Not using generative AI, in future, would be like telling others now that you do not use Google search,” said SUSS’ Cheong.
Such tools must, however, be “regulated to promote transparency, explainability and accountability”, he added.
Not all clients may like the idea of using generative AI. Said Yuen Thio: “Some of our large institutional clients have instructed us that we are not to use generative AI on our work for them due to confidentiality concerns.”
Source: Business Times © SPH Media Limited. Permission required for reproduction.
ADV: D2E: Data Privacy and Protection (5 Feb 2024)
Court dismisses investor’s application for injunction against No Signboard, CEO and controlling shareholder
1
[ad_2]
Source link