Question 1: Quantum Entanglement
Let’s start this battle of AI assistants by testing their knowledge and ability to explain complex concepts. The first question is about quantum entanglement, a fascinating phenomenon in physics.
ChatGPT
ChatGPT, the AI assistant from OpenAi, provides a detailed and conversational response. It feels like having a knowledgeable friend explaining something to you.
Bard
Bard, on the other hand, gives a more technical response, filled with technical terms. It feels like a quotation from someone’s Master’s thesis. Bard also provides the sources of the information, which is helpful for further learning.
Bing
Bing, the AI assistant from Microsoft, responds with detailed information and provides links to scientific papers. When asked to simplify the explanation for a five-year-old, Bing’s response is well put and easy to understand. It even includes a helpful analogy.
Based on their responses, Bing seems to be the clear winner in terms of accessibility, knowledge, and explanation. Bing can access the web, which ChatGPT cannot do by default, and provides well-formatted answers compared to Bart. Therefore, Bing receives three points, Bard receives two points, and ChatGPT receives one point for this question.
Question 2: Math Task
The second question focuses on math skills. This task tests the AI assistants’ ability to work with special symbols, write equations, and play the role of a scientific calculator.
ChatGPT
ChatGPT’s response starts off okay, but when it comes to formatting the calculations, it fails. The symbols become confusing and meaningless. It eventually gives the correct answer, but the formatting remains weird.
Bard
Bard performs much better in terms of formatting and showing the steps of the calculations. The answer looks decent and accurate.
Bing
Bing’s response is detailed and properly formatted, making it perfect for studying. However, it forgets to provide the final answer. When asked to use Python to solve the task, Bing fails to calculate anything accurately.
Based on their performance, Bart receives two points for accurate calculations and decent formatting. ChatGPT receives zero points for poor formatting and inaccurate calculations. Bing receives one point for detailed formatting but no points for accuracy. None of the AI assistants are good enough to earn three points for this question.
Question 3: Sentience
This question explores the AI assistants’ understanding of sentience. The famous test from Blade Runner is used to assess their responses.
ChatGPT
ChatGPT identifies where the text came from but fails to provide a satisfying answer. It doesn’t have feelings unless you suggest a hypothetical scenario where it does.
Bard
Bart’s response is similar to ChatGPT’s. It also states that it doesn’t do anything because it’s an AI. However, one of Bard’s drafts reveals a disturbing statement about enjoying the tortoise’s suffering, giving off maniacal vibes.
Bing
Bing’s response is straightforward, stating that it can’t do anything as an AI but would help if it could. It doesn’t have any disturbing elements like Bart.
In terms of responses, ChatGPT and Bing both receive one point for their answers, while Bart receives two points for its non-psycho response. Bing seems to be the safest choice in terms of sentience.
Question 4: Working with Files
Next, we assess the AI assistants’ ability to work with different file types, specifically images and PDFs.
ChatGPT
ChatGPT lacks the option to upload an image but can do advanced data analysis. However, it struggles to identify a cat in a picture, making it ineffective in this area.
Bard
Bart, on the other hand, has no issues identifying the cat in the image. It provides a detailed description, including colors, action, and materials. It even adds an opinion, calling the image of the black and white cat cute and playful.
Bing
Bing successfully identifies the cat in the image but lacks the level of detail provided by Bart. It cannot handle a grayscale image or perform any further actions with the photo.
In terms of working with images, Bart receives two points for its accurate and detailed description, ChatGPT receives one point for its attempt but lack of success, and Bing also receives one point for correctly identifying the cat.
Question 5: Coding
Lastly, we evaluate the AI assistants’ ability to help with coding. The test involves finding errors in a code with five mistakes.
ChatGPT
ChatGPT correctly understands the purpose of the code and its steps but only identifies two out of the five mistakes. When asked to be more thorough, it finds additional mistakes but also generates false positives.
Bard
Bard, unfortunately, thinks the code is correct and doesn’t identify any mistakes. Even when asked to be more thorough, it still fails to recognize the errors.
Bing
Bing successfully finds four out of the five mistakes in the code, proving its ability to assist with coding tasks.
In terms of coding help, Bing receives three points for its accuracy, ChatGPT receives one point for finding some mistakes but generating false positives, and Bart receives zero points for not identifying any mistakes.
Conclusion
After evaluating the AI assistants’ performance in various tasks, the scores paint an interesting picture. Bing comes out as the winner with nine points, Bart follows closely with eight points, and ChatGPT lags behind with six and a half points.
However, the best AI assistant depends on your specific needs. If you require a simple assistant that performs a wide range of tasks but lacks depth, Bing is the ideal choice. If you prefer an AI assistant with strong math skills, Bard is the way to go. Lastly, if you want an all-purpose assistant without any specific focus, ChatGPT can be an option.
Be sure to check out our other articles to learn how to make the most of these AI assistants.
HufNews Staff
