Sent an AI bot to a job interview in his place – and it didn't end well

An American YouTuber sent an AI agent to a job interview in his place against an AI-based recruiter. The agent fabricated professional experience, offered a salary much lower than what the YouTuber requested, and even scheduled another interview on its own. The experiment raises questions about the level of trust that can be placed in AI agents for decision-making.

American content creator Graham Dot Ziff, a comedian and YouTuber known for AI pranks, created an AI agent named Victor and sent it to a job interview with an AI-based recruiter. During the interview, Victor attributed to Graham skills and professional experience he did not have, and misrepresented his role in a certain project. Graham, who watched the conversation, had to intervene and explain that his representative was making things up. When it came to salary, Graham asked for $140,000–$150,000 per year, but Victor told the recruiter he would be happy to receive only $60,000. The automated interviewer clarified that it was not willing to offer more than $90,000. The most amusing moment came when Graham left the room and left the two agents to manage the conversation themselves; they scheduled another interview without asking him, and then got stuck in an infinite loop of thanks and well-wishes. Graham has garnered millions of views on similar videos. The experiment raises a serious question about the level of trust that can be placed in AI agents to make decisions on our behalf, especially as technology companies develop agents for shopping, negotiation, and closing deals.