AI – What to expect in 2024?


Picture with the help of DALLE


Forecasts are always difficult. In the dynamic world of artificial intelligence (AI), they are particularly challenging.

The rapid development that we experienced with generative AI models in 2023 will also shape 2024.

A recent survey of almost 3000 AI experts indicates an accelerated development.
50% of the experts now expect AI to be able to perform in every possible task better than humans by 2047. A year ago, the experts in the same survey expected this to be 13 years later.

Amara’s law illustrates that the short-term effects of technologies are often overestimated, while their long-term potential is underestimated. This is particularly relevant in the case of AI, as social, entrepreneurial and political adjustments also play a role alongside technological advances. These often require more time.


Picture with the help of DALLE

Knowing all this, I see following developments for the year 2024 and will look at them in more details over the year:

1/ Technological improvements:

Major breakthroughs in new AI models are unlikely, but we will see significant improvements in the quality of models. The focus will shift from Large Language Models (LLM) to Large Multimodal Models (LMM), with open source models becoming more important.

2024 will not bring any major technical breakthroughs in new models. However, we will see many improvements. New training methods, better data quality and fine-tuning will improve the quality of LLM and form the basis for many new applications. The focus will shift from Large Language Models (LLM) to Large Multimodel Models (LMM). These will also become better and better as open source models and OpenAI will no longer be the sole benchmark for the leading models.


2/ Generative AI in companies:

The productive use of generative AI in companies will increase, especially in larger organizations. The challenge will be to identify suitable processes and tasks for the use of AI.

In 2023, only a few companies will have used generative AI productively. Studies have shown that the use of generative AI can significantly increase productivity. In 2024, we will see a significant increase in the use of generative AI in companies. This will be particularly the case in larger companies that have the necessary experts and data. In 2024, it will also become clear that every company will have to address the topic and develop a strategy. More solutions from start-ups and large IT companies will come onto the market, making it easier for companies to adapt. It will be particularly important for companies to understand for which processes and tasks the use of generative AI is particularly suitable – and where it is not. This will not be possible without trial and error, which is a very unfamiliar approach for many companies, especially in Germany.

3/ Impact on the labor market:

AI will lead to productivity gains, which can mean both job losses and the alleviation of skills shortages. A broader approach to training with a focus on AI technologies will be crucial.

The increased productive use of generative AI will have its first impact on jobs and the labor market in 2024, and we will see the beginning of a development that will lead to an increase in productivity and associated job losses through the use of AI. At the same time, the use of AI will help to address the shortage of skilled workers in many countries. However, this requires an increase in knowledge and acceptance of AI among the population and employees. This will lead to a more prepared approach to training with a focus on technologies such as artificial intelligence.

4/ AI agents and marketplaces:

We will see an increase in AI agents and assistants that handle complex tasks autonomously. Products in this area, both for business and private use, will increase.

2024 will see the development of AI agents and assistants. These are AI systems that can interact with each other to perform complex tasks without human assistance. We will see products both in the business sector and for private use. OpenAI already took a first step at the end of 2023 with the GPTs, which are now also being offered on a marketplace. We will use and pay for these agents for activities and, as humans, take over more of the coordination of the various agents – naturally also supported here by AI models such as GPT4.

5/ Open source models:

The relevance of open source AI models will continue to increase. These models will become more powerful and offer new specialized applications, especially for business purposes.

Open source models will continue to grow in importance in 2024. The gap to today’s leading models such as GPT4 will narrow quickly and there will be the first open source models that overtake GPT4. On the other hand, open source will promote many new specialized models and applications, particularly in the business environment. Open source will become the driver of an open innovation landscape in 2024, giving smaller organizations access to the latest models.

6/ Business models of AI start-ups:

AI start-ups must adapt their business models to solve specific customer problems. Standard products and services are increasingly being supplemented with AI functionalities.

2024 will be a challenging year for AI startups. The business models for AI applications will change. It will not be enough to extend a product with OpenAI functionality. In many cases, these solutions will be replaced by GPTs or CoPilots. AI functionality will be offered in all common software products. Examples include Microsoft with its Copilots, Google and Notion. This raises the question for many start-ups as to which customer problem they are solving that cannot also be solved by these standard products or directly by using GPTs. In addition, customers will increasingly ask which manufacturer they are paying for what. Start-ups must clearly position which customer problems they are particularly good at solving and, if possible, provide a benefit-based pricing model for this. Simply marketing AI will no longer be enough in 2024.

7/ AI in edge devices:

Large Language Models (LLM) for local applications on devices such as smartphones will increase. Apple will play a pioneering role here, with data protection and data security becoming important differentiators.

LLMs for local execution, e.g. on smartphones, will become productive in 2024 and find their way into operating systems. Apple will take on a leading role in 2024 and integrate generative AI technology into its hardware and software. The first steps in this direction were already taken at the end of 2023. In connection with this, the issue of data security and data protection when using language models will become an important differentiator in the marketing of solutions. Apple will clearly position itself against Microsoft, Google and co. In 2024, we will see a large number of new AI solutions that run locally on edge devices. In other words, independent of internet access or central models. This will open up a whole host of new applications for companies and start-ups.

8/ Data and copyright:

The use of data and compliance with copyrights will be a key issue. Court decisions and contracts between large AI providers and data owners are becoming increasingly relevant.

The use of data and copyright compliance will be a key issue in AI in 2024. Court rulings on the use of data in the training of large AI models will be handed down this year. One example is the lawsuit filed by the New York Times against Microsoft and Open AI in the USA. Large providers will increasingly conclude contracts with publishers and data owners for the use of their data. OpenAI has started this with Springer Verlag, and others will follow. The issue of copyright protection will also play an increasingly important role in the race between countries for AI supremacy. The discussion about a fundamental revision of copyright protection and an adaptation to the requirements of the 21st century will begin in 2024.

9/ AI regulation:

The regulation of AI is being intensified. Companies will have to deal with new laws and regulations, especially in an international context.

Regulation will be one of the main topics in the field of AI in 2024. Various countries and regions will pass specific laws and ordinances on the regulation of AI this year. Companies will have to deal intensively with the regulatory requirements. In the EU, the final implementation and adoption of the EU AI Acts is expected. Even if it will take another two years before this is actually incorporated into national legislation, companies will have to deal with the topic intensively starting this year. All companies that operate internationally will also have to comply with the regulations in other countries such as the USA, Canada, the UK and China. This will result in high costs and create a new industry for certification and support in AI regulation.

10/ AI and elections:

The role of AI in influencing elections and spreading misinformation is being increasingly criticized. Cybersecurity in the context of AI is becoming increasingly important.

In 2024, almost half of the world’s population will be able to vote in elections. From Taiwan to India and Russia (albeit with very limited voting options) to the EU and the USA, parliaments or presidents will be elected this year. The topic of misinformation and influence in text, words, images and video through AI will play an important role. We will see attempts at manipulation through the use of AI tools. Direct manipulation of election results is also conceivable. The topic of cyber security in the context of AI will become even more important.


Disclaimer: This blog post is written with the help of AI. I use AI as an assistant to generate, translate and optimize the text. The amount of AI used varies depending on the topic and the content. I always curate and edit the text myself to ensure quality and accuracy. The opinions and views expressed in this text are my own.