A video of Tun Dr Mahathir Mohamad in tears has been circulating online. It is fake. And MCMC is now investigating.

Malaysia's communications regulator says the clip of the former prime minister in a state of grief was AI-generated. Nobody filmed it. Nobody consented to it. Yet it looked convincing enough to spread widely before anyone checked.

MCMC has warned that creating or knowingly sharing the video could carry legal consequences. Enforcement action is being considered under existing communications and multimedia laws. The regulator has not confirmed whether a suspect has been identified.

Why this one matters to you

Here is the practical bit: the warning covers sharing, not just making. You do not need to have generated the clip to land in trouble. If you forward something you know is fake, that is on you lah.

So the habit worth building is simple. Before you tap share on a video of a public figure, pause. Ask where it came from. Check whether any established outlet has reported the same thing. A few seconds of doubt costs you nothing.

Part of a pattern

This is not an isolated incident. MCMC describes a growing trend of AI-generated content targeting prominent Malaysian figures without their consent. The tools have become cheap and easy. Fabricating footage of a real person no longer needs a studio or a specialist.

That is the uncomfortable part. Detection and debunking move slower than the sharing. By the time a clip is called out, it has already made the rounds of family group chats over morning teh tarik.

A test for Malaysia

We see this case as a direct test of whether Malaysia can actually police AI-generated misinformation about public figures. The laws exist. The question is enforcement, and whether it can keep pace with how fast this content travels.

For now, the first line of defence is not the regulator. It is you, and the decision you make before you hit forward.

Source: The Vibes

Unknown
Unknown

Continue Reading

Explainers 3 min read

Could AI help stop the poaching of Malaysia's prized gaharu trees?

Researchers and conservationists are exploring whether AI-powered image recognition and satellite or drone monitoring could help detect illegal harvesting of gaharu, the valuable agarwood-producing tree that is heavily poached across Malaysian forests. The idea is to use AI to spot suspicious forest clearing patterns, track known poaching routes, or flag suspicious trade listings faster than manual patrols can manage. The approach is still early-stage, more proof-of-concept than deployed system, and faces real challenges around dense canopy cover, funding and enforcement follow-through. It's a useful example of AI being applied to a uniquely Malaysian conservation problem rather than imported wholesale from elsewhere, and worth watching for pilot results from forestry departments or NGOs.

Unknown
Explainers 3 min read

A viral 'AI hurricane map' turned out to be fake, here's how to spot the next one

A meteorologist has retracted a map showing Hurricane Isaias's projected path after admitting it was generated by an AI tool and never verified against real forecast models. The hallucinated graphic spread widely on social media before being debunked, reigniting concern about how easily AI-generated weather visuals can look authoritative while being completely wrong. Forecasters are now being urged to clearly label AI-assisted graphics and double-check them against official agency data before posting. The incident is a reminder that generative AI tools can produce convincing but false imagery for anything from storm tracks to maps and charts, not just text. If you share or rely on weather graphics online, check the source before trusting it.

Unknown
Explainers 3 min read

If your team builds AI agents with Langflow, patch now, attackers are harvesting credentials

Security researchers at VulnCheck have confirmed a sharp escalation in attacks exploiting Langflow, a popular open-source framework for building AI agents that IBM acquired via DataStax in 2024. This marks the 12th exploited vulnerability found in Langflow this year, with telemetry showing over 50 exploitation attempts detected in a single day in late August. Attackers are using the flaw to harvest credentials from exposed Langflow instances, putting any organisation running unpatched deployments at risk of data theft. Because Langflow is widely used by developers prototyping AI agents and chatbots, the campaign highlights a broader problem: AI tooling is being adopted faster than it is being secured. Anyone running Langflow should update immediately and audit exposed instances.

Unknown