deepfakes

Scammers are increasingly using AI to generate fake faces and voices in fraud schemes, prompting Meta to deploy its own AI defences that analyse billions of signals to detect patterns and remove accounts at scale. The company says it removed 159 million scam accounts in 2025. Meta's Dr. Rafael Frankel discussed this AI-versus-AI arms race, as concern grows over deepfake-driven online fraud.

Researchers writing in The Conversation say deepfake content is distorting this year's Victorian state election in Australia. Their analysis identifies which groups of voters are most susceptible to manipulated audio and video, warning that synthetic media could influence how people judge candidates and parties ahead of polling day.

A short animated explainer from BBC Ideas and The Open University outlines practical habits for verifying online content, noting that deepfakes, voice cloning and synthetic media have made fraud far harder to detect than the era of distorted hands and obvious glitches. Viewers are sharing it as a reminder to check sources, cross-reference footage and stay sceptical of urgent requests.

A public awareness push is warning that deepfakes, voice cloning and other synthetic media have made online fraud far harder to detect than in earlier years, when glitches like distorted hands gave scams away. A short animated explainer from BBC Ideas and The Open University is circulating, offering practical verification habits people can use to recognise manipulated content and avoid being deceived.

AI-generated deepfake fraud is creating new legal liability questions for banks and companies, according to Bloomberg Law. As scammers use convincing fake voices and video to impersonate executives and clients, institutions face growing uncertainty over who bears the losses when payments are authorized by fakes. The report suggests courts and regulators will play a key role in defining responsibility for deepfake-enabled financial crime.

A new analysis from Recorded Future examines how generative AI and synthetic media are changing social engineering tactics used in fraud and identity theft. The report distinguishes between attacks that scale widely and highly targeted schemes, such as deepfake voice or video calls impersonating executives or relatives. It argues that traditional identity verification methods, which rely on visual or vocal cues, are increasingly unreliable as synthetic content becomes cheap and convincing.

Historians are revisiting the Donation of Constantine, a forged imperial decree that medieval church officials cited for centuries to support papal authority over secular rulers. The document, now widely recognised as an eighth-century fabrication, functioned much like a modern deepfake, lending false legitimacy to claims of power until Renaissance scrutiny exposed it.

Axios argues the next phase of artificial intelligence will be defined by AI systems confronting and testing each other, rather than by humans alone. As AI-generated content floods the internet, defenders and attackers alike will be machines, with AI used to detect, counter and exploit other AI. The idea is drawing attention amid growing concern over deepfakes, automated scams and an arms race between generative models and the tools built to catch them.