The next scam message in your inbox may be written, personalised or even voiced by artificial intelligence — and Singapore Police are reorganising around that reality. A 19 September look inside the Singapore Police Force’s new Cyber Command shows officers confronting phishing infrastructure that can be spun up at machine speed, while scammers increasingly use deepfakes, voice cloning and conversational AI to impersonate people, build trust and keep multiple victims engaged at once.
The Cyber Command itself is not new today: it was operationalised on 3 July 2026. What is new is the clearer picture of how its teams are working under one roof, combining investigators, cyber specialists, the Anti-Scam Centre and disruption teams so intelligence can move in minutes rather than through slower hand-offs between separate units.
01 · How AI is changing scams
Police officers interviewed by CNA described three broad ways AI is increasing the pressure on victims. Deepfakes and voice cloning can make impersonation more realistic. Hyper-personalised messages can use information scraped from the internet to make phishing less generic. And conversational AI can help criminals sustain believable conversations with many targets at once.
That matters because modern scams often depend on persuading victims to transfer money themselves. A message that knows your name, echoes a real delivery, imitates a familiar voice or continues a convincing conversation can be harder to dismiss as obvious spam.
Cyber specialists also say criminals can use AI to help write malicious code and look for weaknesses. The practical takeaway is not that every scam is now AI-generated, but that the technology can reduce the effort needed to create, vary and scale criminal campaigns.
02 · From hours to minutes when phishing sites multiply
Courier-delivery phishing is a useful example. Singapore Police have been warning about fake messages that claim a parcel cannot be delivered because of an address problem, then push the recipient towards a lookalike website seeking personal or banking details.
Cyber Command officers told CNA that scam websites can be generated quickly using legitimate online services and automation. To cope with the volume, teams are using AI and machine learning to identify and disrupt malicious sites, freeing specialists to focus on the wider campaign rather than only one URL at a time.
One cybercrime operator said a hand-off that could previously take a couple of hours may now happen in a couple of minutes because relevant teams sit together and can pass context directly. That does not mean every scam site disappears within minutes, but it illustrates the operational goal: collapse the delay between detection, analysis and disruption.
03 · What Cyber Command actually brings together
SPF says Cyber Command provides a full-spectrum response to scams and cybercrime. Its Cyber Operations Centre is designed to detect and disrupt online criminal activity in real time, including using powers under the Online Criminal Harms Act to disable scam enablers. The existing Anti-Scam Centre sits inside that operations structure and continues work with financial institutions on tracing, intervention and fund recovery.
Investigation teams are also intended to work upstream against transnational syndicates involved in scams, ransomware and malware. The command combines uniformed officers, civilian specialists and cyber-trained personnel across investigation, threat hunting and disruption roles.
A separate technology effort with HTX is the National Scams List, built to support automatic real-time sharing of known scam enablers such as bank accounts with partners. The aim is to help banks identify and suspend potentially abusive accounts before stolen money is moved onwards.
04 · Recent operations show the scale of the response
Several September operations give a sense of what this model looks like in practice. From 1 July to 31 August, Cyber Command and digital-payment-token providers identified more than 355 scam victims and prevented over S$8.94 million in potential losses using blockchain analysis and targeted intervention. Intelligence from the operation was also shared with the FBI and New South Wales Police, helping identify 50 foreign victims.
Over the same July–August period, the Anti-Scam Centre and five banks used robotic process automation to disrupt more than 400 scam attempts. More than 3,300 SMS alerts were sent to over 2,700 bank customers, with police saying more than S$46 million in potential losses were averted. Investment and job scams made up a large share of those detections.
Separately, an islandwide operation involving Cyber Command and all seven Police Land Divisions between 27 August and 9 September saw 254 people investigated over suspected links to more than 652 scam cases involving around S$7.5 million in reported losses.
05 · Crypto tracing is becoming part of ordinary scam policing
Cyber Command is also trying to deepen in-house technical capability rather than relying only on commercial tools. A police cybersecurity professional specialising in cryptocurrency investigations told CNA he can support investigators with blockchain tracing, build workflow tools and help identify potential high-value victims.
One internally built tool was used to help pinpoint threat actors attempting to move cryptocurrency through a mixer, a service designed to make blockchain transactions harder to trace. That kind of capability matters because cryptocurrency scams can move money across jurisdictions and services quickly, while the people investigating the crime may need specialised knowledge to follow the trail.
For residents, the bigger point is that “cybercrime” is no longer a separate technical niche from everyday fraud. Phishing, impersonation, investment scams, fake jobs and stolen crypto increasingly overlap, which is why the police are building a command that combines financial, investigative and technical work.
06 · What residents should do when a message feels almost real
Better police technology does not remove the need for a pause at the user end. SPF’s current advice is to ACT: Add security features, Check for signs of a scam, and Tell the authorities and people around you about scam encounters.
- Add protection: use ScamShield, enable two-factor authentication on bank, social-media and Singpass accounts, and consider sensible transaction limits.
- Check independently: do not use the phone number or link in a suspicious message to verify it. Go to the organisation’s official app, website or published contact channel yourself.
- Treat urgency as a warning sign: delivery, banking and government impersonation scams often try to force a quick decision before you verify the story.
- Be sceptical of convincing voices and video: deepfakes and voice cloning mean “it sounded like them” is no longer enough on its own. Confirm unusual money requests through a second channel.
- Tell quickly: contact your bank immediately if money may be at risk, report scam encounters through ScamShield or Police channels, and warn others who could receive the same campaign.
The ScamShield Helpline is 1799. Anyone with information for Police can also use the Police Hotline at 1800-255-0000 or the official i-Witness channel.
07 · The bottom line
AI is giving scammers tools to make old fraud patterns faster, cheaper and more persuasive. Singapore’s response is increasingly built around the same principle: use automation to detect campaigns at scale, combine technical and investigative specialists, and move information to banks, platforms and overseas law-enforcement partners quickly enough to interrupt the money flow.
The 19 September insight into Cyber Command is therefore less about a single new gadget than a shift in operating model. As scam campaigns become automated, police are trying to make disruption more automated too — while still relying on residents to slow down, verify and report.
08 · Sources and live checks
This report was checked against CNA’s 19 September 2026 look inside Cyber Command, the Singapore Police Force’s 3 July Cyber Command announcement, and SPF’s September operational updates on bank intervention, crypto-scam disruption and the latest islandwide enforcement operation.
Last checked: 19 September 2026. Scam tactics change quickly. For current warnings and reporting options, use the official ScamShield and Singapore Police channels rather than forwarded messages or unofficial advice.