Artificial intelligence (AI) has advanced rapidly in recent years and promises to revolutionize several sectors of society. However, what to expect from AI in 2024? In this post, we will explore the trends and prospects for this incredible technology. Get ready to discover how AI will continue to transform our lives and impact the near future. Let's go!
The last year was marked by the popularization of generative AI, with terms like ChatGPT and Bard becoming popular. We were also able to see huge investments in the area, such as the US$10 billion that Microsoft put into OpenAI [1] and the US$4 billion from Amazon into Anthropic [2].
While researchers and CEOs debated the likelihood of Artificial General Intelligence (AGI). Artificial General Intelligence) on social media and newspaper and magazine headlines [3], policy makers have begun to take AI regulation seriously – the European Union has introduced the most comprehensive set of policies governing the technology to date [4] and the The Biden administration published an Executive Order detailing 150 area requirements for federal agencies [5].
If you're interested in AI, 2024 promises to be as busy as 2023. Far be it from me to play fortune teller, but there are some trends worth keeping an eye on over the next 12 months. They can impact both the development of the area and people’s daily lives. These are topics that I will stay tuned for and would like to share with you. Let's go to them:
What to expect from AI in 2024: GPU shortage
Among the AI trends for 2024 that I am very concerned about is the global shortage of processors GPU – the video processors that run most of the AI models we see out there. As many of the large companies that are already in the process of implementing the technology in their businesses have been trying to bring AI capabilities “in-house”, there is a certain race for GPUs. There are several manufacturers (such as NVIDIA and AMD), but their production capacity may be reaching its limit [6].
It is possible that this situation will create enormous pressure not only to increase GPU production, but also for researchers to present hardware solutions that are cheaper and easier to manufacture and use. There is some work in electrical engineering on low-power alternatives to current GPUs, such as that of Stanford researchers Kunle Olukotun [7], who intend to develop specialized and highly efficient chips, built for specific purposes instead of current ones, which can be applied to a wide range of tasks, and Kwabena Boahen [8], which focuses on reducing the heat produced by processing, improving energy consumption.
This research is still a long way from being made available on a mass scale, but I believe there will be enormous pressure to accelerate these efforts in order to democratize access to AI technologies.
The next generation of generative AI
Regarding this topic, I will pay attention to two things. One of them is the emergence of artificial agents and the ability of AI applications to connect to other services, such as search tools. If 2023 was the year we could talk to an AI, in 2024 I believe we will see AI agents do things for us. I'm talking about restaurant reservations, travel planning, among other services.
Additionally, we need to think about what to expect from AI in terms of multimedia. This is the other trend I referred to at the beginning of this section. The new wave of generative AI will be multimodal, capable of understanding and processing multiple types of input, including text, voice, images and even melodies. You will be able to generate all forms of content simultaneously. For example, she could write an article and generate an accompanying image or create a scenario for a play while generating scripts for that same play in several languages. These enhanced skills open windows to more creative opportunities and wide-ranging experiences.
What to expect from AI in 2024: Increased work capacity
Another trend we could see in 2024 involves leveraging AI to increase value in our work processes. AI has the potential to improve human capabilities in several ways, including streamlining processes, increasing efficiency, and maximizing safety. For example, it could be employed for tasks such as content summarization, research, document writing, and code generation and testing [9].
This would promote the development of specialized AIs for each industry, be it legal, medical or software development. It certainly presents opportunities for companies to create specialized AI solutions that align with augmented work trends, but it also brings challenges.
In this sense, we will need to give ourselves time and space to articulate what we consider to be permitted and where we should draw the line. For example, in May 2023, academic journal publisher Springer issued a statement in which it stated that large language models could be used in writing articles, but would not be admitted as co-authors in any publication [10]. The reason they cited, and I think it is extremely relevant, was accountability.
This doesn't mean that Springer is stuck with this understanding forever, but it shows why the idea of accountability is so critical: taking something seriously, understanding what its foundations are, and saying this is where we are now makes clear how we understand the situation at the moment and how we could add more nuance to that understanding in the future. I think institutions and organizations should take this perspective and try to put on paper guidelines for their use of AI in 2024.
Ethical AI
The growth in the use of AI across sectors raises concerns about transparency, fairness, potential job destruction, and uncertainty surrounding the control of AI itself.
In 2024, there will likely be a greater emphasis on ethical AI, focusing not only on ethics but also on fairness and security. Ethical AI is based on principles that prioritize individual rights, security, privacy and data protection. These values guide the development and deployment of AI systems to ensure they do not harm individuals or society as a whole [11].
Developers are likely to come under greater pressure next year to focus on creating socially responsible AIs whose use actually contributes to human growth. Their employers will also feel this pressure for ethical AI, whether through legislation or demand from their customers, but the requirement for them to demonstrate that their AIs are qualified to be considered ethical will increase.
AI Legislation
As AI continues to develop and be increasingly used in various economic sectors, many countries will need to accelerate legislation to control AI. As I mentioned at the beginning of the text, the European Union reached an agreement, marking its first regulation on AI [4]. China, on the other hand, is already beginning to implement laws to restrict the generation of Deepfakes without consent [12].
The challenge for AI legislation lies in finding a balance between protecting people from AI-related harms, such as job destruction and data privacy, and making room for the growth of technology-driven innovations and markets.
For example, in mid-2024, two North American states – California and Colorado – are expected to adopt regulations addressing automated decision-making in the context of consumer privacy [13]. While these regulations are limited to AI systems that are trained by or collect personal information from individuals, they both offer consumers the right to opt out of having their data used by AI systems that use it to classify their profile, e.g. , when contracting insurance.
This type of legislation will force companies to consider the impact of customers exercising their rights, especially en masse. What might happen if a large company uses AI to assist in its hiring process? What if hundreds of potential candidates request that their data be deactivated? Will it be necessary to have humans review candidate resumes? Does including AI guarantee a better process than what already exists? We are just beginning to grapple with these issues. I believe that in 2024, we will closely monitor the evolution of this type of legislation.
Quantum AI
As we advance into the realm of quantum computing, integration with artificial intelligence (AI) promises to revolutionize our data processing capabilities. Quantum superposition and entanglement, the foundations of this technology, have the potential to accelerate AI algorithms in unprecedented ways. However, we face significant challenges such as quantum decoherence, which threatens the stability of quantum information, and the complexity of developing reliable and scalable quantum hardware.
Despite these obstacles, optimism regarding quantum AI is justified, given its ability to perform calculations that would be impractical on classical computers. Continued research and development in quantum AI is essential and we expect to see significant advances in the coming year. The intersection of quantum mechanics and AI is an exciting frontier that promises to expand the boundaries of what is possible in terms of computing.
Looking ahead to 2024, we anticipate even greater integration of AI into everyday applications. Smarter chatbots and AI-powered imaging tools are expected to become ubiquitous, from search engines to social media platforms and enterprise software like Microsoft Office. We are on the cusp of an era where AI will not just be a useful tool, but an essential extension of our computational and creative capabilities.
POR 
ENG