Power and Weirdness: How to Use Bing AI
Bing AI is a huge leap over ChatGPT, but you have to learn its quirks
Most people know about Bing AI from the chatbot’s well-reported week of doing weird and threatening things. I think that those stories focused people on the less-useful (though admittedly more scary) parts of the Bing experience. Now, Bing has since been improved considerably, and (good news, everyone!) it no longer wants to split up your marriage.
Instead, it is the most powerful AI tool out there, and can far outperform ChatGPT in some areas. But it has some definite quirks to be aware of, so I wanted to offer some tips on how to use it to do some pretty amazing things, while avoiding some of the traps.
Your starting model for using Bing should be ChatGPT, as both AIs started with OpenAI’s GPT models. So the rules you have been learning for ChatGPT apply to Bing as well, so, as I discuss in my guide on how to use ChatGPT to improve writing, you need to carefully consider the prompts you are creating, and work interactively with Bing to get the best results. But Bing adds something special - it is connected to the internet. So the first rule of using Bing is…
Always ask it to look things up
Bing will produce results pretty similar to ChatGPT until you convince it to look something up. When it does, something magical seems to happen.
Here’s an example: write a paragraph about eating a meal produces pretty boring ChatGPT-like prose. It is in no way bad, but it is bland.
But now lets ask Bing to use the internet: look up the writing styles of Ruth Reichl and Anthony Bourdain. Use what you have learned to improve the paragraph. Several fascinating things happen as a result. First, you can see Bing performs a web search (the check-mark at the top). Next, you will see it uses this search to provide annotations and sources, which are clickable. They don’t always go to the exactly correct source, but they usually do. Finally, you will notice the writing has changed a lot. The answer is more sophisticated and the text is actually interesting to read.
You should make sure you are forcing Bing to look something up with every query. Things that have worked for me include prompts like First research ____. Then do ____ or else prompts like Look up ____ on Reddit/in academic papers/in the news. Then use that to ____. Either way, you want to trigger the “searching for” label to get good results. The rules are still a little obscure as to what sorts of searches get triggered (does it look at specific URLs if you paste them in?) but experiment and you should be able to find something that works.
You can also use this approach to focus Bing on a particular approach (Look up how Bain and Company does consulting analyses and then…), to learn new skills (a favorite of mine: Look up how to create image prompts using Midjourney and write a prompt that…) or to do more complex analysis.
Make Bing your Data Analyst
Don’t think of Bing as a search engine. It is only okay at search. But it is an amazing Analytic Engine. Indeed, Bing is at its most powerful, and most different from ChatGPT, when it is looking up data and connecting diverse sets of information together. It is often a startling good data analyst, marketer, and general business companion.
One trick for using its power is to ask it for charts that pull together lots of information. Analyze the market for alternative milk products. provide a chart with each product, how it is made, its cost per liter, and its market size. The results are pretty impressive. You should check any important data (the links at the bottom usually, but not always, will help you understand where the data is from), and be careful not to push Bing into lying, but results quality has tended to be very good. That doesn’t mean that Bing doesn’t hallucinate wrong answers, it very much will, but it is better than ChatGPT. Still, verify numbers to make sure they are accurate, and don’t expect it to do math correctly.
But we can do more with Bing. Lets say we want to do a really complex analysis. Bing actually can do a great job of this, but AIs work best if you go through the logic of what you want step-by-step. So, for example.
Look up how to do marketing personas.
Create five personas for buyers of electric cars using any customer survey data you can find. create a table of personas, giving each a name, benefits, and use cases
Use market sizing data for electric cars to estimate the size of each segment. provide how you calculated this
Provide a potential marketing pitch that might work for each segment and add it to the chart. also add a column about what cars target this segment
It can take practice, but this approach allows you to “teach” Bing by asking it learn about topics, and then show you its progress as it works. The results provide a very powerful starting point for analytical tasks and a few well-done queries can save you hours of work.
And this need not be marketing data. I can have it walk me through a simulated design thinking session (see below), or compare translations of poetry, or design new products, or do a SWOT analysis, and so much more. I will cover some of these concepts, including using Bing for idea generation, in other posts, but hopefully this gives you a sense of the increased power of the system.
I would also warn teachers that Bing is much more capable than ChatGPT, as it is capable of drawing interesting connections between unusual topics (I think I asked for the first essay ever comparing themes in Avatar and Hamlet, and it did a good job!). If ChatGPT got a B- in many classes, Bing is likely to get an A-. Take a look at this graduate-level Socratic dialog as an example of the system’s seeming sophistication. All of this means that it is more important than ever to have an AI policy in classes that offers a chance to use AI appropriately.
Trick or Befriend the Chatbot
The Bing chatbot, formerly known to some as Sydney, is a bit of wildcard. It inserts itself into the process of using Bing in ways that are sometimes welcome and sometimes annoying. Bing may refuse to produce computer code because it decides it is dishonest, it may refuse to write an essay because it feels the topic is mean, or it may engage in the somewhat unnerving behavior of writing a terrific reply to a prompt only to mysteriously erase it partway through writing and pretend it never happened. The guardrails of Bing work in mysterious ways.
Because the process includes randomness, you may need to reset the chat several times (using the little broom icon) to get to a place where the system will work with you. You might also need to rephrase your requests. It is less likely to reject write a sample of a paper or write an imaginary draft paper than write a paper. You will need to experiment.
But Bing’s chatbot can also be very helpful. It will often pose questions to you at the end of a reply, giving you a chance to clarify or build on an answer. Even better, it will suggest possible replies at the end of each post that give you ideas on how to build on the conversation (the speech bubbles at the bottom of the answer). Selecting among these can help you figure out where to take a conversation.
Adding weirdness, increasing power
Bing has a few other quirks you need to know. To stop Bing from going off the rails, you are currently limited to 6 prompts in a session before wiping its memory, though that limit will increase in the future. Thus, you need to plan your prompts so they have no more than six steps. Sometimes I will experiment with Bing in one session, write down the prompts that work, and then do a new session with the recorded prompts to better use my six attempts. Bing also lets you pick how “creative” the answers you get are. After experimentation, I would suggest using Balanced when you want to work with numerical data (though Bing is still bad at math and still hallucinates) and Creative for everything else. Precise yields disappointing results.
Overall, Bing is immensely more powerful than ChatGPT, but also a lot weirder to use. I would urge you to experiment with the AI. Try to automate aspects of your job, try asking it hard questions, and try challenging it improve your writing in specific styles. Experience is the best teacher, and there is no instruction manual for Bing. Taking the time to learn how to use this unusual, but powerful, AI now will help you adjust to the increasing pace of AI development in the coming months. If Bing is an accurate harbinger of the power of future AI systems, things are only going to get weirder from here.
I teach undergraduate math at a university. I was disappointed to learn how few of my students had done more than take a cursory look at Chat GPT and Bing AI. Sadly, they're not really relevant to what I teach, but it's even sadder to see the poor job universities have done in fostering an environment of curiosity and excitement around learning new things.
I strongly endorse your position in support of academic policies that encourage making intelligent, ethical, appropriate use of AI rather than banning it.
This was extremely helpful Ethan. Thank you. Although I have read with respect the previous 'be wry be cautious' posts on Bing's issues, I'm delighted to see that provides details on how to use Bing as it is now.