In 2024, AI is poised to become more useful and accurate than ever

Periodic financial statements recently posted by Wall Street companies show that the companies with the best results were Meta (Facebook), Microsoft and Amazon, three technological giants, thereby proving to investors that they not only talk about artificial intelligence (AI), but actively integrate the innovative technology into their products and services. The dramatic impact of AI, and particularly generative AI, was highlighted at the World Economic Forum’s annual meeting in Davos. Representing AI21 Labs, I was proud to present the new developments in AI coming from Israel and to stand in line with the world’s biggest companies. The ideas discussed at Davos provided insights into the likely directions and challenges in the AI sector in 2024 and beyond.

First and foremost, the upcoming year stands to be a turning point for many organizations planning to transition from curiosity about GenAI to the implementation and integration of AI systems. The majority of Fortune 500 companies have already begun experimenting with artificial intelligence, and we have seen an increase in the assimilation of this technology in recent months. This trend is expected to increase in 2024 and the year after, making a significant shift towards AI integration across various industries, from banking, e-commerce, and service sectors.

Ori Goshen.

The expected surge in the integration of generative AI will also enhance the reliability and accuracy of existing models. Many CEOs have been hesitant so far to implement the technology due to concerns about the reliability and accuracy of AI-generated results. Consequently, in 2024, we will see a shift towards the development of more reliable and accurate models.

Businesses are likely to move away from integrating large and general language models (Large Language Models or LLMs), which, in principle, are capable of performing a large number of tasks but are also costly and carry a higher risk of producing hallucinations or fabricated answers. Instead, they will prefer to invest in specialized and smaller models (Small Language Models or SLMs) that are tailored to the specific needs of the business. Similar to the general models, these specialized models, drawing information from a diverse range of sources, are designed to provide more focused answers. For instance, banks will seek out a language model that is task-specific to provide data in response to financial queries, where accuracy is critical. The transition to smaller and more specialized models is therefore expected to cut costs and improve the accuracy and reliability of the AI tools.

Another expected trend is the move from AI built on models to AI built on systems. This means that we can expect to see AI systems that are pre-designed to offer a broader range of responses than current models, such as providing access to organizational information. In this way, the new systems will act to compensate for at least some of the issues inherent in current language models, as previously mentioned. In this context, we can expect to see a step up in the type of AI-generated responses. Rather than generating responses that merely summarize information, we will see systems engaging in complex reasoning - providing analysis and drawing conclusions to help users make informed decisions.

Having said that, the outlook is not entirely rosy. The advancement of AI also involves potential dangers. For instance, malicious actors could create artificial content to spread fake news, sway the opinions and decisions of consumers and voters, and potentially manipulate democratic elections. Additionally, the language models in question have the capacity to mimic human behavior, thereby facilitating sophisticated cyber-attacks. Artificial intelligence also raises ethical questions concerning copyright and its societal impact on fields such as education and higher education. These issues are justifiably capturing the attention of more and more regulators globally, who are exploring ways to increase oversight and measures to mitigate the risks involved.

At the same time, 2024 is expected to herald an alleviation of some of the apprehensions linked to artificial intelligence. For example, the apprehension of robots displacing human workers and leading to mass layoffs is giving way to the recognition that human employees will remain essential and their work will be optimized thanks to artificial intelligence. In fact, AI is set to be gradually integrated across virtually all professions, and will facilitate making the work smarter. This development will further the positive trend of technology democratization witnessed in recent years. Through this trend, AI enables small businesses to augment their capabilities and reach the awareness of thousands of customers and consumers, just like huge corporations. Small businesses, often challenged by the costs of financing marketing campaigns themselves and who do not have public brand recognition, can leverage AI tools to break into new digital markets. This will consequently lead to a more decentralized and equitable market landscape.

This is the reason for the escalating rise in global investments in AI. Talks about a potential "bubble" in the industry do not deter investors, and even at Davos, there was no topic hotter and more popular than AI. It is no wonder that regulatory efforts are intensifying to bridge the gap between current regulations and rapid technological advancements. This is evident in the groundbreaking law recently adopted by the European Union (the AI Act) to comprehensively regulate the field for the first time, including addressing potential cyber threats to AI of the kind mentioned earlier. Therefore, the risks associated with artificial intelligence are not going away, and some are even expected to stay and increase. However, the ability to use AI in an efficient, accurate and more supervised and controlled manner is expected to improve in 2024.

Finally, Israel can and should play a central role in the AI revolution. Governments worldwide are eager to engage in this sector and are determined not to fall behind. For example, during my meeting in Davos with the Hong Kong Minister of Finance, he emphasized how important it is to them that the world's leading AI companies become acquainted with and recognize the potential of the local HK market. Hong Kong is not alone. The knowledge and expertise coming from Israel can therefore remind the world that we are a start-up nation, especially in today's challenging times. This could create advantages for Israel that are not only technological and economic, but also political and strategic.

The writer is a co-founder and co-CEO at AI21 Labs.