Can AI-generated Estonian speech be used to scam people in Estonia?

Despite repeated warnings, Estonian residents continue to fall victim to phone scammers. While the Estonian language has historically protected native speakers from fraud, experts warn that advances in AI voice technology could change this. Investigative programme 'Eyewitness' tested whether AI-generated Estonian speech can convincingly deceive people.

Estonian residents continue to fall prey to telephone scammers despite widespread public warnings, and a new investigation by the programme 'Eyewitness' () has put AI-generated Estonian speech to the test. The question at the heart of the experiment: can artificial intelligence produce convincing enough Estonian to trick real people?

Until now, the Estonian language has served as a natural barrier against phone fraud. Estonian is considered an exotic and highly complex language, making it difficult for foreign scammers to operate effectively among native speakers. However, as AI voice synthesis tools become increasingly sophisticated, that protection may no longer be reliable.

Experts have warned that the rapid development of AI technologies is lowering the threshold for criminals looking to exploit language-based trust. Voice cloning and real-time speech generation tools are already capable of producing passable imitations of various languages, and Estonian is no exception.

The 'Eyewitness' editorial team conducted a live test to assess just how believable AI-generated Estonian speech sounds to ordinary listeners. The results raise serious questions about the future of phone fraud in Estonia and what new precautions residents may need to take.

Security specialists stress that public awareness remains one of the most important defences against scammers, regardless of the technology they use. As AI continues to evolve, Estonians are being urged to stay sceptical of unexpected calls, even when the voice on the other end sounds familiar or speaks perfect Estonian.