The arrival of ChatGPT and its recent integration with Bing has reignited the competition between the search engines.
After Google accidentally presented Bardhis chatbot artificial intelligence search engine, the leading company in search engines seeks to refine it.
To this end, it is essential that the same workers of Google train Bardwho gave an inaccurate answer to a question asked during a demonstration of it Google.
Bard’s mistake
During the demonstration, a user asked Bard “What new discoveries from the space telescope James Webb can I tell my 9-year-old son?”. The artificial intelligence replied with a series of bullet points including one that said “JWST took the first images of a planet outside our own solar system,” a misinformation.
According to the NASAthe first image showing an exoplanet (any planet outside our solar system) was taken by the European Southern Observatory’s Very Large Telescope almost two decades ago, in 2004.
Google training Bard
Google last week ordered its workers to spend between two to three hours helping to train Bard.
Jack Krawczyk, indicated in an internal communication, that the company believes that adding “high quality” answers dramatically improved the quality of the artificial intelligence.
Krawczyk has asked the workers to ask Bard on topics in which they consider themselves experts. Faced with a “very human” answer that is wrong, inaccurate, or doesn’t make sense, they will have to rewrite it to improve a Bard.
