Highlights from OpenAI's Latest Releases
Two weeks ago, Microsoft-backed OpenAI published a blog post unveiling updates to their flagship product, ChatGPT. The former "chatbot" is now being relaunched as an assistant with enhanced capabilities. It not only understands and generates text but also possesses the ability to see, hear, speak, and assist in generating images with the new DALL-E 3. ChatGPT also receives another long-awaited update - internet access. In this article, we summarize some of the highlights from OpenAI’s latest releases.
1. ChatGPT can now see, hear, and speak
ChatGPT is now equipped with image and audio features, allowing for a more dynamic interaction beyond text. With the new audio capabilities, users can now communicate with ChatGPT by speaking directly to it, receiving responses in audio format. Additionally, users have the flexibility to choose from five different voices for their assistant.
For instance, you can now talk to ChatGPT about the latest trends in the electric car market while you walk to work, or have ChatGPT read a bedtime story to your child when your imagination falls short. These new functionalities are powered by OpenAI's speech recognition system, Whisper, originally launched in September 2022.
ChatGPT’s new audio and image features are gradually rolled out to users. The audio functionalities are already available to ChatGPT Plus and Enterprise users on iOS and Android platforms. The image features will also be launched for ChatGPT Plus and Enterprise users across all platforms.
2. Access to the internet (again)
OpenAI has re-enabled internet access for ChatGPT via Bing. OpenAI’s CEO, Sam Altman, wrote, "We are so back" on the platform X, previously known as Twitter, when OpenAI announced the news. Earlier this year, the browsing feature was available to ChatGPT Plus and Enterprise users but was withdrawn due to concerns about users bypassing paywalls.
Now, with internet access restored, ChatGPT can assist users in tasks such as summarizing current news or finding the best offers on products, like an Apple Watch. Additionally, users can see where the information provided by ChatGPT was retrieved from.
Currently, the browsing feature is exclusive to ChatGPT Plus and Enterprise users, but OpenAI plans to extend access to all users soon. To use the browsing feature, select "Browse with Bing" under "GPT-4" in the ChatGPT chat window.
3. DALL-E 3 💜 ChatGPT
OpenAI’s text-to-image generator, DALL-E 2, has been upgraded to DALL-E 3. The initial version of DALL-E was launched in January 2021, followed by the second version in September 2022. Now, a year later, DALL-E 3 is released, promising substantial improvements in understanding context, nuances, and details. Due to the upgrade, it will be better at creating images that more accurately match text inputs (prompts). It also excels in generating images that include text (see example below).
DALL-E 3 is also equipped with additional safeguards to prevent the creation of inappropriate or hateful images. Additionally, DALL-E 3 will not be able to create images of public figures based on their names or mimic living artists’ styles.
DALL-E 3 will also integrate with ChatGPT, meaning that users will be able to leverage ChatGPT to craft perfect prompts for DALL-E 3, making ChatGPT a “prompting partner” and freeing users from having to master Prompt Engineering. If the generated images don't meet your expectations, you can adjust it through text input with ChatGPT.
The new version of DALL-E is already available in Bing for Bing Chat and Bing Image Creator users and is going to be available to ChatGPT Plus and Enterprise users in October.
From chatbot to assistant
As the new capabilities of ChatGPT enables users to interact with it in new ways, it also signifies a pivotal advancement in ChatGPT's evolution, transitioning from a mere chatbot to a multifaceted assistant. Additionally, the new voice and audio features positions ChatGPT alongside Siri, Alexa and Google Assistant - an advancement few saw coming.
What key takeaways can professional users and organizational decision-makers extract from OpenAI's latest releases? The new features not only enhance ChatGPT's functionality, boosting task efficiency for individuals in new ways. It also paves the way for transformative organizational-wide change.
Perhaps your company faces challenges that need solutions involving text-, audio-, and image capabilities? For example, automating your customer support process may require a system that not only can understand text, but also analyze images effectively. If that’s the case, it might be worth looking into how ChatGPT, with its new functionalities, can be a valuable asset in your customer support workflow.
More news to come
Follow Violet on LinkedIn to stay updated on the latest developments in AI.
Violet AI was founded in 2018 as one of the first pure-play AI agencies in the Nordics. Today, Violet consists of a fast-growing consulting and advisory team and four subsidiaries with a total of 50 employees. We specialize in Machine Learning, Advanced Analytics, Intelligent Automation, System Development, and AI Strategy.