2023 is the year where “Artificial Intelligence” or AI has dominated the discourse. Much of this has been dominated by OpenAI’s ChatGPT, the chatbot which has gone viral since it launched in November last year, and already has over 100 million users. Microsoft has also gone all in on AI and invested $10 billion in OpenAI along with announcing a new version of Bing which will be integrated with the start-up’s AI technology. In response to this Google has also announced its own chatbot called Bard, which has had a rough start thanks to one incorrect answer. Let’s take a look at key developments in AI news in the past few weeks. Google Bard gets a disastrous start Bard is supposed to be Google’s own AI chatbot, the answer to ChatGPT if you will. It is built on LaMDA-Language Model for Dialogue Applications, which the company showcased back in 2021 at I LaMDA is a conversational chatbot just like ChatGPT but Google has kept it in limited testing. Now it looks like that threat caused by ChatGPT has forced Google to showcase its own tech in AI. Bard is supposed to be built a limited model of LaMDA, which Google says will ensure less computing power is needed. Bard is right now only accessible to ‘trusted testers’, but Google will soon open it to public testing as well. However, the exact date and time have not been specified. Bard’s introduction to the world didn’t go as planned. Google put out a promotional tweet and video showing off Bard’s capabilities, but one of the answers it gave about the James Webb Space Telescope was inaccurate. Bard claimed the telescope was the first to take a picture of an ‘Exoplanet’ which is not true. The subsequent error has caused Google parent Alphabet’s shares to fall drastically with $100 billion of the company’s market value being wiped out. Alphabet’s shares saw their most massive fall in three months thanks to this ‘error’. Clearly, it also shows that the market is convinced that Google has somewhat lost its edge in AI, and has a chance of losing out to Microsoft in the search market- one which it has dominated for decades now. Google bringing generative AI results to search In the middle of all the hype around Bard, this part of Google’s announcement has not gotten a lot of attention. In his blog post, Google CEO Sundar Pichai also wrote that search results will also include results in ‘easy-to-digest’ formats. In the post, Pichai wrote that these answers will help gain users a deeper perspective on complex queries. He wrote, “Soon, you’ll see AI-powered features in Search that distill complex information and multiple perspectives into easy-to-digest formats, so you can quickly understand the big picture and learn more from the web: whether that’s seeking out additional perspectives, like blogs from people who play both piano and guitar, or going deeper on a related topic, like steps to get started as a beginner. These new AI features will begin rolling out on Google Search soon.” Once again Google is relying on generative AI (or AI which is capable of generating text in this case) to give these results. The information will be sourced based on what is available on the web and its accuracy will be closely watched. Microsoft’s Bing search and Edge browser get the power of OpenAI Just a day before Google held its ‘AI event’ in Paris, Microsoft announced deeper integration of OpenAI’s large language models into Bing Search and the Edge browser. For one, Bing Search is getting a chat feature, where it can answers in a conversational style- just like ChatGPT does at the moment. The difference is that Microsoft says they are using a larger model than the one powering ChatGPT, and presumably, this will mean Bing will be capable of handling more queries. Microsoft claims their AI tech will ensure Bing is more accurate, and faster when giving these answers in the chat. Bing will even write poems, short stories, etc for users, help them plan a project or a party, give ideas for a three-course meal, etc. But Microsoft is also warning that not all of the queries will be accurate and users are being advised to rely on their own judgement as well. Microsoft also showcased how the Edge browser will get a chat and compose feature in the sidebar as well. Chat will let users get a summary of the webpage or document they are reading and even ask questions about the content. The ‘compose’ feature will act as an assistant of sorts helping write text, email, social media posts, etc. Microsoft to offer ChatGPT in Azure Cloud Earlier in January, Microsoft also announced how it planned to offer ChatGPT to its enterprise clients as part of the Azure OpenAI service. Azure already offers some key OpenAI AII services to select customers, and ChatGPT will be added to this list as well. The chatbot has been trained and runs inference on Azure’s AI infrastructure as well. “ChatGPT is coming soon to the Azure OpenAI Service, which is now generally available, as we help customers apply the world’s most advanced AI models to their own business imperatives,” Nadella announced. ChatGPT gets a paid version OpenAI also announced a paid version of ChatGPT just a week back. The paid version costs $20 and is called ChatGPT Plus. Users who sign up for this will be able to access the service even when there is an overload- which is a common problem on ChatGPT right now. They will get a faster response time and priority access to new features and capabilities. This is a pilot right now and only select users can sign up for this. The company had also announced plans to launch an API for ChatGPT where other businesses could plug the chatbot into their own services. More details are awaited on this.
