Welcome to the 40 new deep divers who joined us since last Wednesday. If you haven’t already, subscribe and join our community in receiving weekly AI insights, updates, and interviews with industry experts straight to your feed
DeepDive
Your weekly immersion in AI.
OpenAI has hinted at an imminent release date for ChatGPT-5, with a blog post in May confirming that the new model was already in its training stage. The announcement sparked a widespread hunt for information about what this more mature version of the trending GenAI tool will offer to users.
So how will ChatGPT-5 be different from earlier iterations?
Expanding the capacity of accessible GenAI
Sam Altman (CEO at OpenAI) said in a statement, “Right now, GPT-4 can reason only in extremely limited ways, and its reliability is also limited.”
So the purpose of GPT-5 is to strip away some of those limitations and improve on its existing functionality.
Multimodal capabilities will be a major feature
OpenAI is working to implement multimodal capabilities that enable GPT-5 to generate speech, images, code, and video content – not just text.
This will mark a significant shift in the model’s capabilities compared to previous versions, making it useful for a much broader range of creative and professional tasks. The goal is to make GPT-5 highly versatile as a tool for users – allowing them to use it for multimedia content creation, advanced processing of auditory and visual data, and assisted programming.
It’s an important change for developers, content creators, and marketing professionals
ChatGPT and other generative AI models are already being leveraged in creative industries – by technical creatives (such as developers), professional content creators, and artists. The extended capabilities of GPT-5 could increase its impact on these sectors in a big way, with the potential to transform the way content is created and interacted with.
Of course, there are caveats to that; across industries, we’ve already learnt that the hype around a new GenAI release is usually dampened (at least a little) when businesses actually start using the tool. Instead of an instant radical transformation, we’re seeing a slower, steadier kind of change; as professionals learn to integrate AI tooling into their working practices and then manually improve on its outputs, rather than handing over their creative work to AI completely.
OpenAI is striving for customisable GenAI
The ability to customise and personalise the tool and its outputs is also a focus for the new version of ChatGPT.
It’s expected to offer more tailored responses to individual users – drawing on data from their interaction history and contextual information in their input prompts to deliver results that are more relevant to them.
This has the potential to dramatically improve AI-powered customer service, digital tutoring, and personal assistant applications; as the model learns its users in more detail and gathers information about what will (and won’t) be helpful to them.
And aiming to improve the accuracy of results
ChatGPT is known for its hallucinations: the inaccurate, or sometimes even completely fictional responses that it delivers to users.
By doubling down on improvements to its deep learning algorithms, and training those algorithms on more diverse datasets, GPT-5 is likely to offer more reliable results. And this will make it a more valuable tool for use cases including accurate content creation, and customer support – providing a better user experience overall.
Increased accuracy will, in part, be because of its enhanced ability to understand context. Instead of taking each new prompt as it comes, GPT-5 will be better able to ‘remember’ information from earlier in the interaction, offering increasingly coherent responses as the interaction continues.
OpenAI has a clearer focus on ethics
Importantly, OpenAI has continued to strengthen its focus on developing ethical AI, so we expect to see ethical advancements in this new iteration of its GenAI model.
As of May 2024, OpenAI Board has created a Safety and Security Committee, headed up by CEO Sam Altman, and directors Bret Taylor, Nicole Seligman, and Adam D’Angelo.
The committee will be “responsible for making recommendations to the full Board on critical safety and security decisions for OpenAI projects and operations,” and its first task was to spend 90 days evaluating and developing the company’s safety processes – with ChatGPT-5 at the forefront of their planning.
When the committee was announced, a statement said:
“OpenAI has recently begun training its next frontier model and we anticipate the resulting systems to bring us to the next level of capabilities on our path to AGI. While we are proud to build and release models that are industry-leading on both capabilities and safety, we welcome a robust debate at this important moment.”
The process of developing safe AI practices and ethical boundaries will be slow and winding. But in the short term, we expect to see more effective mechanisms in GPT-5 for preventing the generation of harmful content, and for navigating sensitive subjects.
If you want to keep up with the constantly evolving AI landscape, pre-register now to attend DeepFest 2025.