Deepfakes Become More Terrifying: Zoom Officially Releases Human Detection Feature

The online meeting platform Zoom has officially partnered with World, a human identity verification company owned by Sam Altman, to ensure that meeting participants are truly human and not the result of artificial intelligence (AI) manipulation. This step is taken amid the increasing threat of deepfakes in business communications.

One of the most dramatic cases occurred in early 2024, when the engineering company Arup lost approximately $25 million USD. The case began when an employee in Hong Kong unwittingly approved a fund transfer after attending a routine online meeting with the CFO and several colleagues.

However, it turned out that all participants in the video call, except for that employee, were AI-based deepfakes. Similar attacks were also reported to have affected multinational companies in Singapore in 2025.

Losses due to deepfake-based fraud continue to rise. In the first quarter of 2025, total losses are estimated to exceed $200 million USD.

The average loss per corporate incident now even exceeds $500,000 USD, according to industry security reports. This makes deepfakes a serious threat, especially for companies that routinely conduct high-value transactions via video conferencing.

Previously, efforts to detect deepfakes in meetings generally relied on video frame analysis to look for signs of manipulation. However, this approach is considered increasingly ineffective because advancing AI technology makes visual manipulations harder to detect.