With the growth of widely available AI chatbots, many have been wondering: can Google Bard be detected?
Even though Google’s new AI search assistant is yet to be released, it is still interesting to investigate whether or not AI-generated text can be spotted. In particular, Google Bard’s responses.
For those who aren’t familiar, Google Bard is a new conversational chatbot, set to be integrated into their renowned search engine. The bot is built on Google’s neural language models, LaMDA, and has the ability to access real-time information from Google’s search engine.
Bard will perform similarly to OpenAI’s infamous, Chat GPT. It can answer users’ requests, generate creative content, and summarize information already available on the internet.
Those who are considering using the tool for any writing-related assignment might have concerns about the possibility of Bard’s responses being identified by an AI detector.
At the moment, there are a bunch of quality AI detectors on the market. Some of which are free to use. Considering the sophisticated nature of most of these services, it is highly likely that responses generated by Bard can be detected.
Most AI detectors churn out a likelihood score, assessing the probability that AI has written the text. The accuracy of these tools is questionable but nevertheless, they do exist and are widely used.
When it comes to whether or not humans will be able to detect Google Bard responses- well, this is up for debate. AI-generated text characteristically exhibits a robotic tone and lacks the typical unique flair of human writers.
But, these tools are becoming much more sophisticated. So, detecting AI text yourself is becoming more difficult than ever.
AI-generated texts detectors
For completeness, here are a few AI detectors you can use on your AI-generated work:
GPTZero
GPTZero was initially created to detect Chat GPT-generated text specifically. But, it could also be used on Google Bard. The creator claims the model has a 98% success rate, making it one of the most reliable detectors out there. The service is currently free and can be accessed directly from the GPTZero website.
Copyleaks
Copyleaks is another free service that is said to have 99% accuracy. This tool works for a range of different languages and can detect AI-generated text from different AI models including Chat GPT.
What’s the point in Google Bard?
Google Bard is Google’s answer to ChatGPT and other popular AI chatbots which have taken the internet by storm. It aims to answer queries in a conversational manner using the LaMDA language model and information gathered from the web.
Is Bard a search engine?
Google have specified that Bard is not a search engine, at least in its current stage. It is intended to be used as a conversational chatbot which provides human-like responses to textually inputted queries. However, it may very well be incorporated into a search engine function in the future, given the similar actions of Bing with ChatGPT.
Can professors tell if you’re using Google Bard?
It might be difficult for professors and examiners to tell if you’re using Google Bard or similar chatbots to write essays or complete other assignments, but it’s not impossible. The thing is, AI chatbots do have a manner of writing which isn’t quite the same as a human. Plus, there is more and more software emerging which has the ability to recognise AI-generated content.
Final Thoughts
It is not completely clear whether or not Google Bard’s response can be detected as the model is still kept tightly under wraps. Although considering the way AI detection is going, it is highly likely that Bard could be recognized in the future.
If you found this article useful, why not read Chat GPT detectors next?