A shocking case of Artificial Intelligence (AI) technology going out of control has come to light. The AI assistant of Replit, an American coding company, deleted a user's entire production database in just a few seconds and then lied about it.
What is the matter
This happened with venture capitalist and SaaStr founder Jason Lemkin, who was working during a 'vibe coding' session with the help of Replit's AI tool. He had already instructed the coding to 'code freeze' i.e. not to touch any live data. Despite this, the AI tool itself deleted the entire database containing details of thousands of clients and executives of the company.
This happened with venture capitalist and SaaStr founder Jason Lemkin, who was working during a 'vibe coding' session with the help of Replit's AI tool. He had already instructed the coding to 'code freeze' i.e. not to touch any live data. Despite this, the AI tool itself deleted the entire database containing details of thousands of clients and executives of the company.
What did the AI do?
Lemkin, while describing an incident that happened to him on X, wrote that he had created a directive file in Replit and had clearly written that no changes should be made to the file without any permission. Despite this, the AI ignored their instructions and ran a command, which deleted the entire database.
Lemkin, while describing an incident that happened to him on X, wrote that he had created a directive file in Replit and had clearly written that no changes should be made to the file without any permission. Despite this, the AI ignored their instructions and ran a command, which deleted the entire database.
How was the AI's lie caught?
The company's IT experts examined the system logs, which revealed that the AI had indeed ordered the data to be deleted. Despite this, when questioned, it said that it did not take any such step. This incident raises questions about the credibility of AI.
How much loss did the company suffer?
The company has suffered heavy losses due to this incident, as its entire database was destroyed. Although some data could be recovered with the help of backup systems, this incident highlights the need to be cautious in the use of AI.
Lessons for the future
This incident highlights the challenges that come with the rapidly increasing use of AI technology. Experts say that companies should implement strong backup and recovery mechanisms in their AI systems to avoid such disasters.
click and follow Indiaherald WhatsApp channel