Does it work? Yes, it actually gives you responses with citations. Microsoft has unveiled what is potentially the future of search with it’s conversational Bing search engine that uses ChatGPT. In case you are unaware, there is an ongoing AI arms-race across big tech companies with potentially risky outcomes. For example, Google’s attempt at demonstrating their AI based search Bard, it made a mistake erasing 100bn dollars off shares. That doesn’t seem to bother Microsoft with a 1bn dollar access to OpenAI, the supposedly non-profit organisation behind ChatGPT.
New AI-based conversational search in Bing
In recent years, the field of artificial intelligence has undergone a significant transformation with the rise of large language models. These models have redefined what is possible in the realm of natural language processing (NLP) and have led to unprecedented advancements in the ability of machines to understand and generate human-like text.
In essence, a language model attempts to predict what sequence of words are likely to come next given a sentence. For example, they might generate “little star” given “twinkle twinkle” because it’s likely those phrases co-occur in that order.
Yet, the ability to model what comes next seems to be powerful enough to open a new era of natural language based AI. The concept of language models is not new and people have been researching it for a long time. Some of the very early models date back to the 1950s. But, the rapid growth in computational power over the last few decades and data availability has allowed for the creation of much larger neural network models, leading to the current generation of giant language models such as OpenAI’s GPT-3 (Generative Pretrained Transformer 3).
However, it is important to note that the rise of large language models is not without its challenges. For example, there are concerns around the ethics and fairness of AI models, particularly in regards to the potential biases that can be learned from the data they are trained on. There are also concerns about the energy consumption required to train and utilise these models as well as the carbon footprint of the AI industry as a whole.