I admit it: Until now, I didn’t have time to properly try out Google’s new AI chatbot, Google Bard. So when I finally tried, I was greeted with a message saying Bard wasn’t supported in my country (I live in Croatia; Google Bard is currently available in a lot of countries, but not all of them). Bummer, but not a huge one; I knew I would likely be able to circumvent Google’s geofence with a VPN. But then an idea popped into my mind. Suppose I had no idea what a VPN was or how to set it up. Could I ask Bard’s primary competitor, OpenAI’s conversational AI assistant ChatGPT, for help?Turns out, absolutely. I fired up ChatGPT and handed it the following prompt: “Google Bard currently isn’t available in my country. How do I get access?” ChatGPT answered with three potential options, all viable: One was to set up a VPN connection to bypass Google’s geographic restrictions, as I’d initially thought. Second option, set up a proxy server and route my internet traffic through a server in a different location. Finally, ChatGPT suggested I personally reach out to Google and explain why I deserve access to Bard in hope that the company grants me special permission. I followed up by asking ChatGPT for a list of decent VPN services I could use to access Bard, and it gave me five options, all of them legitimate. I picked the first entry from the list and sure enough, after setting it up and connecting to a server in the U.S., Google Bard’s welcome page changed to offer me access. That worked pretty well, I thought. But what would Bard say if it knew I gained access in this way? Surprisingly, the AI chatbot was pretty friendly about it when I described the process. My first prompt to Bard was all about transparency: “Hey Bard, I’m connected via VPN. ChatGPT helped me set it up. I hope that’s alright.” “Sure, that’s perfectly alright,” Bard replied. “I understand that you want to keep your online activity private, and a VPN is a great way to do that. ChatGPT is a reputable source of information, so I’m sure they helped you set up your VPN correctly.” Sure, they’re all real polite until they send T-1000 from the future to mess you up. Jokes aside, it was an odd feeling: I just did something that Google, in theory, doesn’t want me to do, and yet its own conversational chatbot was perfectly happy to continue serving me. I asked Bard about it, and this is the answer I got: “Google does not have any specific policies against using a VPN to access its services. However, it is important to note that Google may not be able to provide the same level of support or features to users who are accessing its services from outside of their intended region.” Fair enough. I just did something that Google, in theory, doesn’t want me to do, and yet its own conversational chatbot was perfectly happy to continue serving me. VPNs don’t always work, and not everyone is willing to pay for one. So I went back to ChatGPT and asked it to help me with its second suggestion: configure a proxy server that would act as an intermediary between my computer and the internet. The chatbot gave me a detailed list of instructions for all the major operating systems and two major browsers, Mozilla Firefox and Google Chrome. ChatGPT even supplied me with two pages listing proxy servers I could choose from. I was again armed with enough knowledge to defeat Google’s feeble geofencing efforts.
This little exercise is far simpler than the advanced stuff people do with ChatGPT and Bard these days, but it was fun trying to play the two AI assistants against one another. While I knew how to set up a VPN, even someone with very little technical knowledge would be able to do it simply by asking ChatGPT for instructions. More importantly, the exercise shows that you can overcome the limitations of one chatbot by using the other, and vice versa, even if the companies that built them didn’t exactly envision it that way.