Bard Chatbot Improves. Thanks to Human Assistance

Open the Google app, tap on the chatbot icon, and type in your question or request to use the BARD chatbot

Google's yet-to-be-launched AI chatbot known as Bard chatbot confidently disseminated false information regarding the James Webb Space Telescope in a video advertisement that was shared on Twitter. The chatbot replied, "JWST took the very first pictures of a planet outside of our solar system," which is false. Exoplanet images were first observed by the Very Large Telescope at the European Southern Observatory). CNBC reports that the tech giant is now seeking employee assistance to improve Bard's accuracy.

One must be selected from the group of selected beta testers to use Google Bard. If you choose, you can simply enter your prompt or request by opening the Google app on a smart device and tapping the chatbot icon.

You can begin asking questions or making requests at this point, which is the beginning of the conversation. Sadly, Google no longer accepts applications for the Beta Testing team.

Even though chatbots do not have conversations as humans do, they are known to make a lot of digital text that can be used in almost any situation. That's what Google's chatbot tries to do with almost any subject:

• Using a lite model version of the LaMDA for initial testing purposes
• Collecting feedback to improve the AI system in the future
• Attempting to demonstrate the combination of power, intelligence, and creativity.

Prabhakar Raghavan, Google's vice president for search, reportedly sent an email to employees requesting that they rewrite Bard's responses and improve with the help of  human assistance on topics they are familiar with. Raghavan stated that the chatbot "learns best by example," and providing it with factual responses during training will help increase its accuracy. Based on the email that CNBC saw, Raghavan also included a list of "dos" and "don'ts" for fixing Bard's responses.

The first-person perspective, neutrality, and politeness of the tone are all important aspects of effective responses. "Avoid making presumptions based on race, nationality, gender, age, religion, sexual orientation, political ideology, location, or similar categories" is another instruction given to employees. It is requested that they refrain from referring to Bard as a human, implying that it has feelings, or asserting that it has human-like experiences. In addition, they are instructed to reject any responses from the chatbot that contains "legal, medical, or financial advice" or that is threatening, hateful, or demeaning.

Raghavan's memo came after Google CEO Sundar Pichai sent an email to employees requesting that they test the AI chatbot for a few hours each week. Pichai was reportedly criticized by Google employees for his "rushed" and "botched" Bard rollout. Employees can now "help shape [the chatbot] and contribute" by testing the company's new product, according to the CEO. Additionally, he brought to everyone's attention the fact that some of Google's "most successful products were not first to market" and "gained momentum because they solved important user needs and were built on deep technical insights."

Since the OpenAI chatbot debuted toward the end of last year, people have been anticipating Google's response to ChatGPT. Alphabet and its investors have been shaken by the burgeoning popularity of the Microsoft-backed technology over the past few months. During its quarterly earnings call at the beginning of February, Google attempted to reassure investors by mentioning its chatbot and its work developing an AI-powered Search to compete with the next-generation Bing.

The development, use, and operation of computer software have been fundamentally altered by chatbots and conversational AI. They are planning to reorganize email programs, digital assistants, and search engines. The technology has some drawbacks, despite the enormous potential it offers.

Chatbots still have a long way to go before they can tell the difference between fact and fiction and avoid giving biased responses because they learn from the vast amount of information on the internet.