Having been immersed in the mobile industry for over a decade, I’ve witnessed an enduring rivalry between iPhone and Android enthusiasts. Both sides staunchly defend their preferred platform, unwilling to be swayed. Interestingly, despite the ever-narrowing feature gap between iOS and Android, distinctions still exist, with each operating system boasting its own unique strengths. However, for those unswayed by brand loyalty, it’s worth noting that choosing between iOS and Android today guarantees a stellar mobile experience, regardless of the selection.
Notably, Google’s Bard AI recently expressed a preference for iOS over Android, a stance that, while not as detrimental as Bard’s previous factual mishap that momentarily impacted Google’s stock, is less than ideal from a marketing standpoint for Google.
It’s worth mentioning that Bard’s inclination toward iOS appears to have changed within a matter of days. In my own tests, Bard demonstrated a preference for Android, that is, if Bard could even hold an opinion.
Bard represents Google’s response to OpenAI’s ChatGPT and has been available in a limited capacity since early May. During this time, Google has progressively made the chatbot more accessible, removing the waitlist for users. However, it’s important to note that Bard remains unavailable in the European Union region due to its stricter privacy regulations compared to the rest of the world.
Nevertheless, users can access Bard from anywhere worldwide using a VPN service and pose a variety of questions to it. For instance, many users recently queried Bard about its favorite mobile operating system—iOS or Android.
The findings of these inquiries revealed Bard’s apparent inclination. Surprisingly, Bard did not favor Google’s own Android, which is rather unexpected considering its association with Google’s expanding repertoire of applications.
Bard, to the surprise of many, consistently favored iOS and provided reasons why Apple’s operating system surpasses Android in terms of simplicity and security. This occurrence was not isolated, as Apple Insider was able to replicate the response. It is worth noting that Bard is not a confirmed user of either mobile OS, at least not to our knowledge. However, this preference for iOS expressed by the rival ChatGPT may have some connection to the training process of these large language models.
These models extensively consume data from the internet, including a vast array of articles and reviews discussing iOS and Android. It is plausible that Bard’s inclination towards iOS represents a sort of weighted average of opinions and reviews found online. Alternatively, it’s also possible that Bard has been manipulated or influenced in some way, affecting its preference for Apple’s mobile OS over Google’s Android.
A more effective approach for engaging Bard would be to instruct the bot to provide supporting links that back up its reasoning. This could offer additional insights into Bard’s decision-making process.
However, it is important to note that developments have occurred in the days following the initial tweet. It remains unclear whether Google intervened to correct Bard’s previous responses, although such intervention would not be surprising. What I discovered during my own attempts to replicate the aforementioned answers is that Bard’s stance had changed.
Upon posing the question in various ways, Bard provided different responses. Initially, it claimed it could not have a favorite, which is a reasonable response to expect from a generative AI chatbot following such a prompt.
In the subsequent response I received, Bard expressed a preference for Android, providing a detailed list of pros and cons associated with the operating system. However, when I specifically asked Bard to provide sources and links to support its claims, the bot failed to deliver the requested information in each instance.
Furthermore, during a third attempt where I requested sources and links right from the beginning, Bard provided the same “error” response as the first try, stating that it is a language model and does not possess the necessary information or capabilities to fulfill the request.
These instances indicate that Bard, as a language model, may have limitations in terms of providing specific sources or links to support its claims.