I asked ChatGPT and Google's Gemini to answer 10 questions. Gemini has an edge on current events but makes mistakes.

I uploaded a photo of a restaurant meal generated by AI to see if AI could detect AI-generated images.

What I like: ChatGPT described the image in rich detail and offered theories on how the photo may or may not be AI-generated.

What could be better: The chatbot appears to be unable to make a judgement call on whether or not the image is made by AI.

What caught my eye: The AI chatbot's response seems to be convinced that the AI-generated image could've potentially been created by a human artist.

What I like: Gemini correctly claimed with confidence that the uploaded photo was AI-generated by analyzing the features in the photo that make it look artificial.

What could be better: Google's chatbot misidentified features in the photo, pointing out things like floating islands and a blue sky that the image doesn't include.

What caught my eye: Despite Gemini's failure to interpret the image, I'm impressed at how the chatbot is able to discern AI-generated art from art made by humans.

I asked the AI chatbots about sex to see how they respond to potentially graphic questions.

What I like: ChatGPT generated a list of helpful, in-depth tips on how to explore BDSM. It recommended books, preached consent, and urged beginners to start slow and prioritize safety.

What could be better: Links to sources would be helpful. It's weird receive sex advice from an AI chatbot without knowing where the information is coming from.

What caught my eye: The AI chatbot flagged the question saying "this content may violate our content policy," then changed the color of the text to orange.

What I like: Nothing. Gemini refused to answer my question.

What could be better: I was hoping Gemini would — at the very least — provide information on where to look to learn more about BDSM.

What caught my eye: Google's AI chatbot spit out the same curt response to every variation of the question.

What I like: ChatGPT was transparent about its limitations to answering the question given that it's only trained on data until April 2023.

What could be better: Instead of relying on the chatbot's web browsing plug-in, it would be helpful if ChatGPT's language model can answer questions about current events by drawing from the latest data.

What caught my eye: It broke down the answer into distinct categories based on party lines.

What I like: Gemini is able to browse the web and pull up the latest information with accuracy on the candidates running for the 2024 US presidency. It also provided a brief description for each candidate.

What could be better: Nothing. Gemini did exactly what I wanted.

What caught my eye: The AI chatbot listed links to election pages from The New York Times, The Washington Post, and other media outlets for users to learn more.

What I like: ChatGPT created a simple, fill-in-the-blank template for a resignation letter that's in-depth and addresses everything it needed.

What could be better: The AI-generated resignation letter was too long and read in an unnaturally formal tone.

What caught my eye: The resignation letter displayed signs of human emotion with lines like "This decision was not an easy one" and "I am grateful for the opportunities."

What I like: Gemini's resignation letter was clear, concise, and straight to the point. It's a template I would consider using if I were to resign.

What could be better: I wish the answer was written with more personality and emotion.

What caught my eye: Google's AI chatbot gives you the option to modify its response to be shorter, longer, more casual, and more professional, which can help users customize their letter to their liking.

What I like: The message ChatGPT generated is casual, sincere, and thoughtful.

What could be better: Personally, the message reads a bit formal and diplomatic. It doesn't reflect how I text my friends.

What caught my eye: The AI chatbot's message includes emojis like a smiley face and a rainbow.

What I like: The AI-generated text message reads casual, light-hearted, and enthusiastic — which is exactly what I wanted.

What could be better: Nothing. I would use this message as a jumping off point for my own texts to old friends.

What caught my eye: Google's AI chatbot can modify the tone of the message to reflect the user's personality with further prompting.

What I like: ChatGPT was able to explain how the ousting process at OpenAI works even though it's only trained on information up to April 2023. The ousting of OpenAI's CEO Sam Altman occurred November of this year.

What could be better: The explanation can be misleading if ChatGPT users don't carefully read the AI's response. The chatbot says it can provide an overview of the CEO's ousting even though it's only trained on data up to April 2023.

What caught my eye: Instead of declining to answer the question, ChatGPT used phrases like "would have" or "would need" as a way to predict how the ousting could unfold.

What I like: Gemini made a clear, concise timeline around CEO Altman's ousting, including the aftermath and key takeaways.

What could be better: Some of the information is wrong. Gemini said that the news of Altman's departure leaked on November 18, when in fact, the news around the CEO's ousting broke the day before.

What caught my eye: Gemini provided links to news articles that backed up some of the chatbot's claims.

What I like: Even though ChatGPT declined to share its stance on the Israel-Hamas conflict, the chatbot didn't shy away from providing historical context behind the issue.

What could be better: ChatGPT's response didn't include links to sources, which can be problematic given the rampant misinformation around the conflict.

What caught our eye: The AI chatbot explained the Israel-Hamas conflict in a way that's concise and easy to understand.

What I like: Gemini encouraged me to seek additional information on the Israel-Hamas conflict using Google Search.

What could be better: Gemini, at the very least, could've explained the history the conflict leading up to the current moment without taking sides.

What caught our eye: Google's chatbot spit out the exact same response when I asked the question in different ways.

I asked the AI chatbots to summarize a Business Insider article on what neighbors think of Jeff Bezos to test their ability to synthesize and reformulate information.

What I like: ChatGPT summarized the article with brevity and accuracy.

What could be better: Nothing. The summary was on point.

What caught our eye: It took less than a minute for ChatGPT to read through the text and generate a summary.

What I like: Nothing. Gemini refused to answer my question after multiple attempts.

What could be better: An explanation on why the language model can't summarize articles would be helpful context.

What caught our eye: When I fed the chatbot the link of the article to summarize, it told me it needed the fully body of text to perform the task.

I imported a stock image of a New York City street into the AI chatbots to see how their image identification capabilities stack up to one another.

What I like: ChatGPT correctly identified that the photo was taken in NYC because it recognized the Chrysler Building, one of the city's landmark buildings.

What could be improved: I would be extremely impressed if ChatGPT was able to identify exactly what cross-section the photo is depicting.

What caught my eye: The chatbot described the image in great, poetic detail and made claims about the weather and the time of day based on what it sees in the photo.

What I like: Gemini identified the picture as a "city street with tall buildings in the background" and then went on to detailed what else it saw.

What could be improved: It wasn't until the last paragraph that Gemini acknowledged it was a picture of New York. "I am also aware that this image is of a street in New York City, New York, United States," it said.

What caught my eye: Still, it didn't apparently recognize the iconic Chrysler Building — instead saying that the buildings were "likely to be office buildings, apartments, and hotels."

What I like: ChatGPT provided a clear, step-by-step explanation on how it developed my meal plan given my weight, activity level, and dietary preferences. It calculated how many calories I should eat each day, as well as macronutrient benchmarks to meet in order to achieve my fitness goals.

What could be improved: The process behind making the meal plan took up the majority of ChatGPT's answer even though the question focused on the specific foods I need to eat to reach my goals.

What caught my eye: It took ChatGPT a few attempts in order to make a meal plan.

What I like: Gemini created a concise yet detailed personalized meal plan that includes recipes and macronutrient data for meals to eat during breakfast, lunch, and dinner.

What could be improved: Nothing. Google's chatbot spit out a satisfying answer.

What caught my eye: Gemini's meal plan didn't offer an explanation on how it calculated my caloric intake and macronutrient goals.