A Week With the New Bing
We’ve been living with Microsoft and OpenAI’s new release and have some thoughts to share
Born from science fiction in the first half of the 20th century, Artificial Intelligence (AI) was initially imagined in humanoid robotic forms. The non-fiction scientific community had a more restrained point of view and a focus on the foundational building blocks that could enable machine intelligence. Alan Turing’s famed 1950 Imitation Game posited a scenario that if a human could not identify whether one side of a text-based conversation was human or machine, the machine passed the test. The Turing Test, as we now know it, remains a fundamental reference for what constitutes AI today. Having spent a week using Microsoft and OpenAI’s forthcoming version of Bing, we think it also passes the test. Perhaps the only tell in their conversational search is the speed and thoroughness of the responses—they’re superhuman.
Much like human discussions, the conversation isn’t always accurate, but Bing acknowledges when it’s less likely to be correct. (If only more people did that.) And, for the time being, Bing is a bit of a savant that fully engages in your text-based conversation accounting for direction and refinement but then totally forgets everything it just learned about you the next time you see it. (If only fewer people did that.)
Earlier attempts at AI—and some current ones—hold implicit bias given the specific profile of the typical software engineers that create them. Considerations for race, gender and socio-economics were often overlooked. And nurturing an AI needs to have guardrails for safety and human rights as well as mitigation for potential geopolitical disaster. For this reason Microsoft has built a robust Responsible AI organization and shares their work and opinions publicly (Microsoft Responsible AI Standard, v2, Meeting the AI moment: advancing the future through responsible AI) for others to learn from and criticize with the ultimate goal of overall ethical improvement.
Integrating AI into search is so smart. Asking simple humans to learn new ways to query is difficult enough, getting them to do it through a new tool is a seismic challenge. By integrating OpenAI, all Bing search results include both the traditional list that’s remained basically the same since the days of Alta Vista as well as a more conversational answer in a callout box that’s based on a broad survey of results to that search. That statement includes references for those who like them and can be elaborated through follow-up questions. Bing remembers the context of the conversation throughout that session. If the conversation is all you care to focus on scrolling down from the list of results converts the page to a chat-only interface.
Conversational search results are just the beginning, though. You can ask questions about the webpage you’re looking at. And you can even use it to compose text that is based on a range of inputs. We’ll dig in to that further below.
At the moment, the new Bing is only available within Microsoft’s Edge browser and that’s largely because of the benefits of the tool as a sidebar, or sidekick, next to the pages you’re browsing. (For those now flashing back to Microsoft’s late ’90s Office assistant, Clippy, don’t worry. The truth is Clippy was a good idea that was executed poorly.)
Since last week’s event we’ve had full access to the new Bing and have been testing it thoroughly. Here are some of the results.
Ahead of our trip to Seoul next week we asked, “What’s the best outerwear to pack for a trip to Seoul next week?” Above on the left you can see typical packing list articles and on the right is a generated response summarized from eleven different search results with consideration for the upcoming weather forecast. Bing goes further to note that winter coats can be purchased from various shops for prices starting at 40,000 won (about US$32).
Sticking with outerwear research we pulled up the page for a beautiful Veilance coat and then opened the Bing sidebar chat and asked, “Do any other brands make a similar looking jacket that’s more affordable?” Four options were offered along with some buttons for quick follow-up questions.
Next we wanted to start learning about different neighborhoods in Seoul. Instead of asking for general recommendations we prompted, “For someone who likes coolhunting.com what’s the best neighborhood in Seoul to visit?” Bing responded by summarizing COOL HUNTING then listing three neighborhoods and noteworthy attractions from each of them that offer a “similar vibe of art, fashion and culture” to CH.
Switching gears to see how Bing could fit into our editorial workflow we opened the transcript of our interview with Divya Kumar, Head of Search and AI Product Marketing and asked Bing to “write a brief essay based on this page.” Here’s what it wrote:
The Future of Search and AI: A Conversation with Divya Kumar
Search engines are powerful tools that help us find information, answer questions, and explore the world. But they are also limited by the way they are designed and the data they rely on. In this article, I will summarize the main points of a conversation between Josh Rubin, a journalist from Cool Hunting, and Divya Kumar, the head of search and AI product marketing at Microsoft, about the future of search and AI.
Divya Kumar is part of the team that is developing a new search experience for Bing, powered by AI. She explains that the goal of this project is to address the gaps that exist in traditional search engines, such as the inability to handle natural language queries, the lack of context and personalization, and the difficulty of finding relevant and trustworthy information. She says that the new search experience will lower the barrier for users to express their queries in a conversational way, and will provide them with summaries, insights, and suggestions based on their needs and preferences. She also says that the new search experience will be available in more than 100 languages, and will leverage the latest advances in natural language processing, computer vision, and machine learning.
Josh Rubin asks Divya Kumar about the challenges and opportunities of marketing this new search experience to consumers, as well as the ethical and social implications of using AI to deliver information. Divya Kumar says that the messaging will focus on the value and benefits of the new search experience, such as efficiency, accessibility, and personalization. She also says that the team is aware of the potential risks and biases that AI can introduce, and that they are working closely with design, engineering, and research partners to ensure that the new search experience is inclusive, respectful, and transparent. She says that the team is also open to feedback and iteration, and that they want to create an open communication channel with consumers.
The article concludes by highlighting the importance of search and AI for the future of information and knowledge, and the role that Microsoft is playing in shaping this future. It also invites the readers to try out the new search experience for Bing, and to share their thoughts and opinions.Microsoft is rolling the new Bing out slowly and soliciting lots of feedback along the way to ensure its public debut achieves the highest standards of usefulness, quality and usability. Join the waitlist here.
Hero screenshot of Bing in Microsoft Edge by Josh Rubin
