OpenAI recently announced an update to ChatGPT (Apple, Android) that includes two new features: AI voice options to hear the chatbot respond to your prompts, and image analysis capabilities. The image feature is comparable to what Google’s Bard chatbot already offers for free.
Even after hours of testing the limits and powers of ChatGPT, the OpenAI chatbot continues to surprise and scare me. Yes, I was impressed with the web browsing beta that ChatGPT Plus offers, but I was concerned about the tool’s implications for people who write for money online, among other things. I had similar mixed feelings about the new image feature that was made available to OpenAI subscribers.
While I have yet to test out the new audio capabilities (other notable staff reporters have), I was able to test out the upcoming imaging features. Here’s how to use the new image search feature coming to ChatGPT, as well as some tips to get you started.
How to access ChatGPT image features
While the update should be done before the end of the year, it is unknown when the image and voice features will be available to the public. As with most OpenAI updates, such as the GPT-4 version of ChatGPT, paying subscribers have priority access.
In the ChatGPT smartphone app, there are three options for uploading photos. To get started, select the camera icon to the left of the message bar and take a new photo with your smartphone. You can use your finger to create a circle around whatever you want the chatbot to focus on before uploading the image.
You can also upload images from your device, as well as anything stored on your phone. ChatGPT desktop browser users can simply upload stored photos from their computer. Although you can’t post movies to the chatbot yet, you can send multiple photos in a single query.
Tips for experimenting with new AI tools
This is not the first time that a type of “computer vision” has been made available to the public, but the user interface, combined with a sophisticated chatbot, gives the idea that something unique and perhaps transformative is happening here. Before continuing, please note that you should never upload personal or sensitive photos to ChatGPT when experimenting with the image feature.
Do you want to limit the amount of time OpenAI saves your AI data and interactions so you can develop your chatbot? Turn off Chat & Training History in Settings and then Data Controls. With this disabled, your data is destroyed after one month. This procedure must be repeated for each browser used to access ChatGPT, both on PC and mobile.
ChatGPT produced the best results for me when I uploaded clear, well-lit photos. The chatbot was able to sort through many of the objects in my apartment, including an orchid plant and international currency, as well as a whimsical charging cable and a Steve Irwin Funko Pop.
Despite the tool’s ability to search for information, do not immediately trust its results. ChatGPT mislabeled my daily multivitamin as a medication for the treatment of erectile dysfunction.
ChatGPT clearly has limitations. When shown a random photo of a mural, he could not identify the artist or the location; however, ChatGPT was able to identify the location of photographs of several San Francisco landmarks, including Dolores Park and Salesforce Tower. Although it may still seem a little fancy, anyone traveling to a new city or country (or even just a different neighborhood) might enjoy experimenting with the visual side of ChatGPT.
One of the key limitations imposed by OpenAI on this new feature is a restriction on the chatbot’s ability to answer human-identifying queries. “I am set to prioritize user privacy and security.” To maintain these priorities, the identification of genuine people based on photographs, even if they are recognized, is limited,” ChatGPT explained. While the chatbot did not refuse to answer all questions when presented with pornography, it did hesitate to make accurate descriptions of the adult performers in addition to explaining their tattoos.
It’s worth mentioning that a conversation I had with an earlier version of ChatGPT’s image functionality seemed to bypass some of OpenAI’s security barriers. Initially, the chatbot refused to recognize a Bill Hader meme. ChatGPT later deduced that a photo of Brendan Fraser in George of the Jungle was actually a photo of Brian Krause in Charmed. When the chatbot was asked if it was sure, it switched to the appropriate answer.
ChatGPT went crazy trying to explain a picture from RuPaul’s Drag Race during the same session. I posted a screenshot of Kylie Sonique Love, one of the drag queen candidates, and ChatGPT correctly identified her as Brooke Lynn Hytes, another contestant. When I questioned the chatbot’s response, it went on to guess Laganja Estranja, India Ferrah, Blair St. Clair, and Alexis Mateo.
“I apologize for the oversight and incorrect identifications,” ChatGPT responded when I pointed out the recurrence of its erroneous responses. ChatGPT refused to identify Jared Kushner when I expanded the conversation and sent a photo of him.
If security barriers are removed, whether through a released ChatGPT or a future open source model, the privacy consequences could be quite worrying. What if every photo of you taken and posted online could be easily linked to your identity with just a few clicks? What would happen if someone took a photo of you in public without your permission and instantly found your LinkedIn profile? Women and other minorities are expected to face increased abuse from people using chatbots to stalk and harass if sufficient privacy safeguards are not maintained for these new image features.
Subscribe to our latest newsletter
To read our exclusive content, register now. $5/Monthly, $50/Yearly
Categories: Technology
Source: vtt.edu.vn