Amazon continues to experiment with artificial intelligence and conversational chatbots. On February 1st, the company announced the launch of Rufus, a virtual assistant that would help and guide customers when purchasing any product on the website.
However, the truth is that the chatbot has not proven to be very useful, for now. At least that’s what The Washington Post exposes, where it warns that the work of Rufus leaves much to be desired.
According to Amazon, “Rufus is trained on Amazon’s extensive product catalog, customer reviews, community questions and answers […] to answer customer questions about a variety of purchasing needs and products […] and make recommendations based on conversational context”. That’s something.
However, few people can try Rufus and testify to those features. Only a selected group by Amazon has access to the chatbot and its abilities, which are currently nothing out of the ordinary. For example, in one of the cases, the journalist asked Rufus which would be the best cycling gloves to keep hands warm in winter. In the searches, the bot recommended a pair of short-fingered gloves designed for warm climates.
And the thing is, at this point, it is worth asking if this type of tasks can really be replaced by a machine. The capabilities of chatbots like ChatGPT continue to improve by leaps and bounds, and they are very useful for tasks such as creating summaries or writing code, for example.
However, they are not useful for obtaining objective information. They do not represent a significant improvement compared to technologies like Google’s, as they tend to invent data and suffer from hallucinations.