ChatGPT: 5 changes I’d like to see in the near future

The first anniversary of the release of ChatGPT has now passed. Since then, the chatbot has received multiple updates, adding features such as custom instructions, newer models like GPT and GPT-4 Turbo, and more. But competition has been heating up over the past year, with Microsoft Copilot and Google’s Gemini sometimes providing better results than ChatGPT.

So while chatbots are still popular today, I believe ChatGPT could take some inspiration from the competition. To that end, here are five things I hope to see from ChatGPT in the coming months.

Large language models like the one that powers ChatGPT suffer from a major problem: hallucinations. The term refers to the tendency of modern artificial intelligence to generate text that appears reasonable on the surface but is filled with factually inaccurate or outright misinformation. This effect can be greatly mitigated by “connecting” ChatGPT to the Internet, much like you and I do with search queries on Google.

ChatGPT gained the ability to search the internet in early 2023, but the feature has been locked behind a paywall since then. Specifically, you’ll need to pay $20 per month for ChatGPT Plus to add web browsing support. On the other hand, we have chatbots like Copilot, Perplexity AI, and Gemini, which offer this functionality as a standard feature for free.

With this in mind, I would like OpenAI to provide web browsing support to all users, regardless of whether they have an active subscription. The alternative is dangerous because it means ChatGPT could continue to spread inaccurate information and damage its reputation in the long run. I don’t think I’m asking for too much either – OpenAI can continue to lock better GPT-4 models behind its subscription.

While ChatGPT started out purely as a text-based chatbot, it eventually gained the ability to create AI-generated art. It uses OpenAI’s own DALL-E text-to-image model under the hood, which means it delivers excellent results. However, it lacks many of the features you get by using an AI image generator directly.

The only thing ChatGPT can do is create images based on text prompts. You can’t request minor modifications to individual images, unless you don’t mind the chatbot creating an entirely new set of images. You also can’t upload your own photos or images and ask the AI ​​to edit them, even though that’s a feature available in DALL-E. Finally, ChatGPT cannot upscale your favorite images to a larger resolution. One workaround is to use ChatGPT’s code interpreter to perform basic editing (pictured above), but this is just using a programming tool rather than artificial intelligence.

Given that ChatGPT locks image generation behind a paid subscription, I hope these AI art-related features make their way into the chatbot. After all, competitors like Stable Diffusion and Midjourney offer all of these features regardless of your subscription level. Even Samsung’s Galaxy AI software suite now lets you perform AI-based image editing on your smartphone for free.

A few months ago, I detailed ChatGPT’s voice chat feature, which allows you to have lengthy back-and-forth conversations with a chatbot. The feature has finally dropped from the paid tier, meaning anyone can now use it. However, it’s still a relatively unknown feature that most people don’t even know about.

The mode is only available through the mobile app and is currently just a voice extension for the chatbot. Fast forward to earlier this year, and we found code suggesting that ChatGPT would soon let you set it as your default assistant on Android. Unfortunately, Google beat ChatGPT back in February by releasing Gemini, which lets you replace Assistant and even perform basic tasks like setting reminders.

However, for those of us who enjoy the conversational experience of ChatGPT, I hope the feature continues to evolve. I’m also looking forward to chatbots being able to do things like set timers, get news, and control my smart home.

Currently, you need to go into the ChatGPT mobile app and tap the headset icon every time you want to ask a few questions. I know, it feels very dated.

You might think I’m crazy for asking a chatbot to output less text, but hear me out. When you ask ChatGPT a question, most of the time you’re usually just looking for a small piece of information. So why do I have to read a long preface before I get the answer every time I use ChatGPT? I do expect longer responses from Siri and Google Assistant, but ChatGPT takes that to the other extreme.

I believe Microsoft Copilot strikes the perfect balance between verbosity and information. It offers three different chat modes: Creative, Balanced, and Precise, which you can choose every time you initiate a conversation. As you’d expect from their titles, Precise condenses the chatbot’s output into a few sentences, while Creative gives you a more ChatGPT-like response.

ChatGPT has cleaner responses or a different chat mode that would motivate me to use the chatbot more often, although I do hope this doesn’t come at the expense of accuracy.

If you’ve ever used Gemini, you may have noticed that it can read your email from Gmail and even interact with various Google services like Docs, Maps, and Flights. This is the level of cohesion you’d expect from a true digital assistant, but when you use ChatGPT, the experience is completely absent.

Even on the paid tier, you can only use custom GPT features to interact with external platforms like Wolfram Alpha and Kayak. While useful, they require you to go out of your way to enable them and only last for the duration of the chat. It doesn’t help that OpenAI has now discontinued the plugin feature, which brought multiple external services into your chat at once.

I believe the true next generation of ChatGPT will be one that interacts with third-party services on my behalf, without prior direction. Likewise, I think it would be more useful if the chatbot could read my email and calendar (with prior consent, of course). These changes will make ChatGPT feel more personal and give it more autonomy.