What are the different types of Generative AI?

Generative AI (Gen-AI) is a term which we are starting to become familiar with, thanks in part to how it is being used by different brands and companies and how it is transforming the field of content marketing. But what are the capabilities of Gen-AI? And what different types of Gen-AI exist?

Generative Pretrained Transformer (GPT): a language based Artificial Intelligence (AI) that, using deep learning, is able to generate human-like text. This type of Gen-AI works by being able to generate new text based on inputs that it has previously received and which it has been trained on. Platforms such as Neural Text or Grammarly offer users the chance to create texts based on simple sentence prompts.

‍

Chatbots: beyond text creation and translation, Gen-AI can now be used to create advanced, human-like chatbots. ChatGPT launched by OpenAI recently caused a media frenzy thanks to the chatbot’s ability to “followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”

‍

Text-to-image programmes: For many creative industries, programmes such as Midjourney, DALL-E or Stable Diffusion will be able to change in some part how images are created. By being able to build on or even replicate artwork, gaming, architecture and movies, text-to-image programmes are challenging how we create such content. By prompting such tools with human language, the software is able to build a picture that in a matter of seconds.

‍

Video creation: Gen-AI has the ability to support marketing teams in their video-making tasks. From creating videos with an AI spokesperson like Movio does, to editing a video intelligently, the ways that marketers might create videos in the future is changing.

‍

What do these different types of Gen-AI mean for the future?

Gen-AI exists based on the potential it holds to solve a number of challenges and its ability to open up new opportunities for various industries.

‍

Key areas where Gen-AI can intervene:

‍

• Creating new content: text, images or music can all be generated through Gen-AI. Gen-AI can be trained to understand specific sectors, industries and companies meaning that it can understand how to generate content that is in line with brand values, messaging and expectations.
• Improving efficiency: Through automating certain content production, Gen-AI gives teams like the marketing department time-saving solutions. By reducing the time spent looking for an image that doesn’t exist, creating moodboards for a campaign or having to build a social media post from scratch, marketing teams have the opportunity to optimise their work
• Building high-quality content: Gen-AI has now reached a point where it is convincingly high quality. Gen-AI created work has won art prizes, been used in major campaigns by brands such as Heinz and Nestlé and is continuing to break new ground. Integrating such technology into certain campaigns is now a way of keeping up with current trends and guaranteeing a certain quality to your content
• Generating personalised content: Gen-AI works intuitively and instantly based on information it has learnt about certain elements, like for example a consumer’s preferences on an e-commerce site. Gen-AI can therefore generate content specifically adapted to each user in order to provide them with a tailor made experience of a website or mobile application

‍

What are the risks when it comes to Gen-AI?

Like with any new technology, there are reasons to question how it can be used.

Whilst the ability to create endless new types of content has got many people excited, there is concern for some people that this will take away something from those artists and creatives who create original content. Professor of Art, Carson Grubaugh worries that people working within the creative fields of entertainment, video games and advertising risk losing their jobs. Digital artist Grey Rutkowski has also expressed concerns after his style has been copied thousands of times by different AIs. The issue of how such content is copyrighted is therefore one which needs to be questioned.  

‍

Along with concern for jobs and the authenticity of original content, there are also worries about how such content will be governed. With language models able to churn out large quantities of text, there is room for error, false information or misleading content. Indeed, when Meta Galactica aimed to train a Gen-AI model how to analyse 48 million science articles in order to create summarised academic papers, the platform was taken down after less than three days, having published incorrect information.

‍

There is as such room for such technology to be misused, create fake information or even automated troll bots. Filters for certain platforms are not sophisticated enough to remove inappropriate content or systemic biases, something which will likely take time to iron out. If such AI programmes are to become as widespread as predicted, these issues are going to become ones which need to be faced, particularly if large brands and companies are going to invest in such technology.

‍

What does the future look like for Gen-AI?

Gen-AI certainly looks to be a game-changing technology and offers substantial benefits to creative teams.

‍

When used with tool such as a Digital Asset Management platform for example, a Gen-AI fuels content production and distribution, opening up new ways of working on campaigns and streamlining relations with other parts of the business.

‍

From personalised e-commerce websites to avatar-led training videos, Gen-AI looks to have an exciting road ahead of it. With careful governance and restrained use of the technology, many brands and companies can use this technology to their advantage.

‍

It’s worth remembering that Gen-AI works from human-created prompts, with the human language still being the key to setting in motion the creation of different types of content. Whilst Gen-AI looks set to become a useful tool in marketer’s and creative team’s kit, it remains a support for human creativity and work. All that remains is to watch this space and see just how far Gen-AI will take us.

‍

Want to try it yourself? Test Wedia.ai free of charge

‍