The AI might mislead users through its defective hallucinations just like how Google was criticised for its wrong information and deceptive solutions.

The e-commerce platform Amazon.com started a new initiative introducing a new series of iterative processes on the incremental refinement of several of its artificial intelligence products as it seeks to stay ahead of rivals amid continued investor excitement over the technology on Wednesday, Reuters reports.

Since the majority of the giant companies like Apple, Microsoft and Google combat to keep pace in the AI race, the retailer attempts to counter the rivals accumulating Gen AI in the fray. The company announced its vision and improvements at the New York conference.

Also, read| Waiting For A Good Opportunity To Shop: Amazon Prime Day Is Here

Amazon‘s vice president of Gen AI Vasi Philomin stated that the company is set to add additional memory to agents that automatise work to enhance business as it allows each request to build on prior ace. This would further benefit to provide a seamless and more personalised experience when it comes to complex tasks. If the user requests an aisle or window seat on a flight, the updated AI agent could remember each request in lieu.

The company has updated the Q chatbot, which was released last November, to improve the generation of software code. Amazon has also announced enhancements to its Bedrock service, allowing businesses to create applications with a variety of AI models. These updates help in detecting and filtering out inaccuracies, known as “hallucinations,” wherein AI generates misleading or incorrect responses to questions or requests, the report adds.

Also, read| Amazon May Invest Extra 10 Billion Euros In Germany: Report

The AI might mislead users through its defective hallucinations just like how Google was criticised for its wrong information and deceptive solutions. Matt Wood, vice president at Amazon Web said that the new initiative will help to dwindle the occurrence of such delusions for at least 75% of users providing them with meticulous data.