AI-Powered Dating Conversations Making Online Scams Harder To Detect - Selasie Vondee Warns

The Chief Technology Officer of Smart Innovations Ghana, Selasie Vondee, has warned that advances in artificial intelligence are making online romance scams and fake identities increasingly difficult to detect.

Speaking on the Big Convo segment of The Culture Daily Morning Show on 3Music TV and Pluzz 89.9 FM, he explained the growing use of catfishing and chatfishing in online conversations.

He described catfishing as creating a fake online identity to deceive people, often by using someone else’s photograph, while chatfishing involves using AI tools to generate or manage conversations on behalf of a real user.

Vondee said many people now rely on AI to write opening messages or continue conversations on dating platforms, particularly when they struggle to start conversations themselves.

He noted that although AI-generated messages can sometimes be identified through their writing style, newer tools are designed to make responses sound more natural and human.

“Technology is growing. There are tools that can humanise AI responses to make them sound more natural,” he said.

He also warned that technologies such as deepfakes, voice cloning and AI-generated video are making impersonation more convincing, increasing the risk of fraud.

According to him, many dating platforms now require users to verify their identity with facial recognition before creating a profile, although he said some bad actors have found ways to bypass such safeguards.

Vondee identified platforms such as Tinder, Hinge and Bumble as popular dating apps where AI-assisted conversations are becoming more common. He also pointed to AI assistants such as ChatGPT, Claude, Meta AI and Snapchat AI as tools people increasingly use to draft messages.

Looking ahead, he said new AI features are being developed to help users identify potentially risky conversations.

One emerging technology, he said, does an analysis of messages for behavioural warning signs and assigns a red flag score, allowing users to decide whether to continue the conversation.