AI must remain assistive tool not decision maker says former judge
He emphasized that any AI impacting individual rights requires rigorous testing and clear oversight before deployment.
Harmindar stated AI possesses transformative potential for human endeavor but must not replace human judgment particularly within criminal justice systems.
AI should be an assistant, not a decision maker he declared during his plenary session at the ASEAN Law Forum 2025.
He cited a Sabah case where an AI sentencing tool recommended ten months imprisonment with fifty-four percent probability for a drug offence.
The magistrate instead imposed twelve months imprisonment demonstrating human judgment must prevail.
Harmindar chairs the Criminal Law Review Committee and stressed the necessity for proper rules and safeguards.
A clear risk-based approach is essential before adopting AI in judicial processes he emphasized.
Singapore has already issued guidelines for AI use in courts he noted while the Philippines drafts its judiciary AI governance framework.
Malaysia and other ASEAN members must follow suit with their own frameworks he urged.
A framework ensuring responsible AI use requires risk assessments and mitigation measures he explained.
AI use cases must undergo testing for bias, reliability and transparency before adoption he added.
He cautioned that AI's criminal justice application demands vigilance despite being powerful technology.
We must always remain accountable he asserted regarding AI integration into justice systems.
Approaching AI with eyes wide open prepares us to address risks and reap benefits he concluded.
Colin Seow of Colin Seow Chambers provided Singapore's perspective on AI implementation.
The Singapore Police Force uses AI in digital forensics to automatically sift seized devices for abuse material.
This integration significantly improves investigation speed and accuracy he confirmed.
AI deployment extends to Singapore's financial sector through the COSMIC platform he added.
The Monetary Authority of Singapore launched COSMIC last year as a public-private data-sharing initiative.
Major banks and financial institutions participate in this collaborative information sharing platform.
AI analyses transaction patterns within COSMIC enabling proactive suspicious activity detection.
Banks can flag concerning activities across institutions while safeguarding customer confidentiality he explained.
Legislative changes support this initiative allowing information sharing strictly for crime prevention.
The three-day ASEAN Law Forum featured fifty-eight speakers and moderators across fifteen sessions.
Over three hundred participants from ASEAN and beyond attended the expert-led discussions.
Topics covered alternative dispute resolution and commercial law reforms alongside human rights. - Bernama
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Sun
4 hours ago
- The Sun
AI must remain assistive tool not decision maker says former judge
KUALA LUMPUR: Artificial intelligence must remain an assistive tool rather than a decision-maker according to former Federal Court judge Tan Sri Harmindar Singh Dhaliwal. He emphasized that any AI impacting individual rights requires rigorous testing and clear oversight before deployment. Harmindar stated AI possesses transformative potential for human endeavor but must not replace human judgment particularly within criminal justice systems. AI should be an assistant, not a decision maker he declared during his plenary session at the ASEAN Law Forum 2025. He cited a Sabah case where an AI sentencing tool recommended ten months imprisonment with fifty-four percent probability for a drug offence. The magistrate instead imposed twelve months imprisonment demonstrating human judgment must prevail. Harmindar chairs the Criminal Law Review Committee and stressed the necessity for proper rules and safeguards. A clear risk-based approach is essential before adopting AI in judicial processes he emphasized. Singapore has already issued guidelines for AI use in courts he noted while the Philippines drafts its judiciary AI governance framework. Malaysia and other ASEAN members must follow suit with their own frameworks he urged. A framework ensuring responsible AI use requires risk assessments and mitigation measures he explained. AI use cases must undergo testing for bias, reliability and transparency before adoption he added. He cautioned that AI's criminal justice application demands vigilance despite being powerful technology. We must always remain accountable he asserted regarding AI integration into justice systems. Approaching AI with eyes wide open prepares us to address risks and reap benefits he concluded. Colin Seow of Colin Seow Chambers provided Singapore's perspective on AI implementation. The Singapore Police Force uses AI in digital forensics to automatically sift seized devices for abuse material. This integration significantly improves investigation speed and accuracy he confirmed. AI deployment extends to Singapore's financial sector through the COSMIC platform he added. The Monetary Authority of Singapore launched COSMIC last year as a public-private data-sharing initiative. Major banks and financial institutions participate in this collaborative information sharing platform. AI analyses transaction patterns within COSMIC enabling proactive suspicious activity detection. Banks can flag concerning activities across institutions while safeguarding customer confidentiality he explained. Legislative changes support this initiative allowing information sharing strictly for crime prevention. The three-day ASEAN Law Forum featured fifty-eight speakers and moderators across fifteen sessions. Over three hundred participants from ASEAN and beyond attended the expert-led discussions. Topics covered alternative dispute resolution and commercial law reforms alongside human rights. - Bernama
Barnama
5 hours ago
- Barnama
AI Must Remain As An Assistive Tool, Not Decision Maker
KUALA LUMPUR, Aug 20 (Bernama) -- Any use of artificial intelligence (AI) that impacts individuals' rights must only be deployed after rigorous testing and with clear oversight, says former Federal Court judge Tan Sri Harmindar Singh Dhaliwal. He said while AI had the potential to transform human endeavour, it must remain an assistive tool and not a decision-maker, particularly in the criminal justice system. "AI should be an assistant, not a decision maker. In Sabah, where AI sentencing tools were piloted, the model once recommended a 10-month imprisonment with a probability of 54 per cent for an offence under the Dangerous Drugs Act 1952, but the magistrate imposed 12 months instead. "The message is clear, AI can assist, but human judgment must prevail,' he said during a plenary session titled 'Criminal Justice Reforms: Ensuring Ethical AI in Criminal Justice Proceedings' held in conjunction with the ASEAN Law Forum 2025 at the Kuala Lumpur Convention Centre today. Harmindar, who is also chairman of the Criminal Law Review Committee, emphasised the need for proper rules, safeguards and a clear risk-based approach before adopting AI in judicial processes. "Singapore has already issued guidelines for AI use in courts, and the Philippines is drafting an AI governance framework for its judiciary. Malaysia and other ASEAN member countries must follow suit. "We need a framework for ensuring responsible use of AI. This includes an assessment of risks, mitigation measures, and ensuring AI use cases are tested for bias, reliability and transparency before adoption,' he said while cautioning that while AI was a powerful technology, its application in criminal justice should be approached with vigilance. 'We must always remain accountable. If AI is to be used in the justice system, we must approach it with our eyes wide open, ready to address the risks in order to reap the benefits,' he added. Speaking from Singapore's perspective, counsel Colin Seow of Colin Seow Chambers said the Singapore Police Force had incorporated AI into digital forensics, enabling investigators to automatically sift through seized devices for child sexual abuse material and significantly improve the speed and accuracy of investigations.
Free Malaysia Today
6 hours ago
- Free Malaysia Today
Baidu posts quarterly revenue drop as consumption flags
Baidu achieved revenue of US$4.56 billion in Q2 2025. (EPA Images pic) BEIJING : Chinese internet giant Baidu recorded a slight drop in quarterly revenue today, dragged down by a persistent slump in domestic spending as its push into artificial intelligence (AI) accelerates. Beijing-based Baidu, the operator of China's top search engine, generates a significant proportion of its revenue from online ads, making its performance highly susceptible to fluctuations in the country's spending patterns. The firm achieved revenue of ¥32.7 billion (US$4.56 billion) during the second quarter (Q2) of 2025, down 4% year-on-year, according to a statement to the Hong Kong Stock Exchange. Revenue from online marketing during the period was down 15% year-on-year to ¥16.2 billion, the statement showed. China is facing an uncertain economic outlook as cautious consumers navigate a years-long downturn in the property market, high unemployment and trade tensions with Washington. Retail sales – a key gauge of consumer demand in China – grew at a slower rate in July than expected, official data showed on Friday. Following years of tight regulation of the vast Chinese tech sector, Beijing is hoping that recent advancements in AI will provide the spark needed to jumpstart the domestic economy. Baidu also said today that its net profit during Q2 was ¥7.3 billion – a 33% jump year-on-year but down 5% from the previous quarter. The company has invested heavily in AI, placing it in an increasingly competitive race alongside China's other tech giants Tencent, Alibaba and ByteDance. It has also sought to advance its autonomous 'robotaxi' services abroad. Baidu and Lyft announced plans this month to launch robotaxis on the rideshare app in Germany and Britain in 2026, pending regulatory approval. The firm said in a joint statement with Uber in July that it plans to offer driverless cars on the Lyft competitor's app in Asia and the Middle East this year.
