Generative AI was the explosive breakout tech trend of 2023, capturing the attention of not only techies and futurists but mainstream audiences, too, with its ability to create seemingly human-like content.
With tech giants including Microsoft, Google and Meta pouring their apparently limitless resources into the technology, we can expect equally game-changing innovation in 2024.
Generative AI is an incredibly powerful tool that will play an important role in democratizing access to the transformative potential of AI. And I believe everyone needs to be aware of what’s coming in order to carefully consider the impact it will have on our lives.
So here’s my rundown of what I expect to be the key themes, covering new technological breakthroughs as well as societal issues we’re likely to encounter.
Bigger And More Powerful Models
Generative AI applications are as impressive as they are because they are trained on huge datasets. GPT-4, the engine behind ChatGPT, is said to be trained on over one trillion parameters. Other large language models (LLMs) like Google’s PaLM2 and DeepMind’s Gopher are trained on hundreds of billions of parameters. Rumors are heating up that GPT-5 could be around the corner, and we should expect another big step up in terms of size. Increasing the size of the dataset isn’t the only way to make an LLM smarter, but so far, it’s been shown to be the most reliable way of creating improvements. Because of this, it’s the direction I expect to see the development of new generative AI models heading in 2024.
Electoral Interference
In 2024, we will see leadership elections in many countries, including the USA, UK and India. I’m quite confident we will see generative AI used to spread misinformation and disrupt the political process. This is likely to be through the use of ever-more convincing deepfakes, as well as the use of tools such as ChatGPT to generate huge amounts of propaganda and distribute it at scale. Of course, politicians and parties will also use it to generate personalized campaign messages and emails.
Generative Design
Another field where we can expect to see generative AI rapidly adopted is in the design of physical products and services. Generative design is the name given to an emerging breed of tools that allow designers to simply input requirements and available materials and be given blueprints and recipes. Design platforms such as Autodesk are incorporating generative AI functionality, allowing product designers to rapidly generate any number of prototypes as digital twins and testing them in parallel in order to come up with more robust, effective or sustainable product designs.
Generative Video
Video is the preferred media format among the younger generations, so it comes as no surprise that AI toolmakers have been quick to come up with tools that tap into this trend. Creation of video content has traditionally been expensive, mainly due to the need to hire a trained human professional to create, capture or edit footage. With generative video tools, anyone will be able to produce professional-looking video content for a fraction of the price in 2024.
4 Generative Audio And Speech
As with video and design tools, 2024 will be the year when AI-generated audio and speech become as ubiquitous as AI-generated words and pictures have become in 2023. I expect to see the subtly robotic qualities that are still usually detectable in all but the most sophisticated AI-generated speech start to vanish as tools become better at simulating the tones and inflections that make up human speech.
5 Multi-Modal Models
Most generative AI tools, algorithms and LLMs specialize in simulating one “mode” of expression – language, visuals, or sounds, for example. The trend, however, is towards “multi-modal” generative AI. The coming version of OpenAI’s ChatGPT will be capable of understanding and interpreting images, as well as taking voice commands and talking back to us. Meta has also demonstrated a model that can pull together images, text, audio, depth, and inertial data. This is something that will become more common over the next year, and it should soon be normal to be able to talk to an AI about a picture or a video in the same way we can talk to it about text today.
Prompt Engineers In High Demand
Prompt engineers “program” generative AI systems by describing what they want them to do in the terms most likely to lead to optimal results. It’s been described as “the hottest new job in tech” in 2023, and demand for it is likely to grow in 2024. The job has also been called “AI whispering” as it’s about being able to coax the best out of systems that can occasionally seem complex, incomprehensible and intimidating.
Autonomous Generative AI
Autonomous Agents is a term being used to describe a class of generative AI applications that effectively operate themselves by continuously generating and responding to prompts. By doing this, they are able to carry out more sophisticated operations than chatbot-style generative agents, which simply respond to a prompt and then wait for the user to tell them what to do next. One well-known example of an autonomous agent that emerged during 2023 is AutoGPT. In 2024, I expect to see further progress in this field as AI engineers continue the quest for “generalized” AI that can turn it’s hand to any job we set it.
Generative AI-Augmented Apps And Services
By 2024, 40 percent of enterprise applications will come with conversational AI as an embedded feature, according to AIM Research. We see it happening already with developers like Microsoft and Adobe practically falling over themselves to build chatbots and natural language interfaces into their platforms. Popular social messaging app Snapchat adds a generative AI bot to its users’ contact lists that they can use to get answers to questions or just treat as a virtual friend. In 2024, anyone who sells apps is going to be adding chat interfaces to improve engagement and customer experience.
Generative AI in Schools And Education
During the last century, traditionalists railed against the introduction of electronic calculators in classrooms, predicting students would become over-reliant and unable to do basic calculations without them. What actually happened was students became capable of doing far more advanced calculations at a younger age. Calculators then became so universal that the question of whether we could cope without them became moot. We will see similar debates around the use of generative AI by students during 2024. Teachers and lecturers, meanwhile, will grow used to using it to generate notes, reports, lesson plans and course summaries.