The competitive field of AI chatbots just got tougher as Google unveils its new entry, Bard, designed to rival the popular ChatGPT. However, Google does caution users that, similar to its competitor, Bard may occasionally produce invented or "hallucinated" answers.
Though Google has long-established itself as a leader in artificial intelligence and search, Bard serves as an attempt to demonstrate that the tech giant has not fallen behind. Much like ChatGPT, Bard is capable of answering questions and engaging in discussions on an extensive range of topics, sometimes exhibiting humanlike understanding. Users have found Bard helpful for generating ideas or text, despite its occasional inaccuracies.
The rapid success and hype surrounding ChatGPT have prompted Google to take bolder risks in the chatbot market. While the technology behind ChatGPT and Bard boasts remarkable potential, the unpredictability of AI-generated responses requires further development in order to gain an edge in the competitive world of chatbot AI.
Bard and ChatGPT employ AI models that predict subsequent words in a text based on training data gathered from vast amounts of text resources. While this approach is highly effective in simulating human responses, it can lead chatbots to generate inaccurate or hallucinated facts, which may present problems for users seeking genuine information or web searches. Additionally, AI-generated responses may reflect biases present in the training data regarding race, gender, and age.
Google has released this early version of Bard to a select base of testers, with users in the US and UK now able to apply for access. It will be available through a separate web page from Google's standard search interface, providing users with three generated answers to each question to emphasize the experimental nature of Bard. The interface also allows users to give feedback on the chatbot's performance via a thumbs-up or thumbs-down response, with the option for additional written input.
Contributions from Google AI researchers resulted in several pivotal advancements used in ChatGPT, including the transformer machine-learning algorithm that forms the language model behind the chatbot. While Google had first showcased a chatbot utilizing this technology in 2020, the tech giant opted for a cautious approach, particularly after facing criticism surrounding the sentience claims of a language model they were developing.
However, with Microsoft's recent $10 billion investment in ChatGPT creator OpenAI and the launch of a conversational interface for its search engine, Bing, the AI chatbot race is heating up. Chinese search engine, Baidu, has also announced the development of its own chatbot, Ernie Bot.
Last week, OpenAI revealed GPT-4, an enhanced version of the language model behind ChatGPT. In response, Google announced it will make a powerful language model called PaLM available for use through an API, integrating text generation features into its business software, Google Workplace. Furthermore, Microsoft has showcased new features for its Office suite utilizing ChatGPT.
As companies harness AI chatbot technology and integrate it into various industries, products like custom software development and web app development will become even more crucial. No-code and low-code app development platforms, such as AppMaster.io and others, offer an accessible and affordable way to build backend, mobile, and web applications.
