In the past few years artificial intelligence have rapidly evolved making many theoretical concepts like note-taking a practical reality of AI-based transcriptions. ChatGPT has proven that AI can surpass human abilities in certain areas ushering in the age of superintelligence. This article presents the top AI trends that are going to cause a stir around in 2024.
Agentic AI
According to Jill Goldstein, global managing partner of HR and talent transformation at IBM Consulting “We’re entering a new chapter in how employees get work done with the rise of AI agents. Unlike AI assistants, AI agents can generate plans based on a prompt and carry out tasks independently.” Agentic AI, designed for autonomous decision-making, will take over routine front-line tasks such as responding to customer inquiries, scheduling, and lead qualification. By the end of 2025, AI agents will independently manage customer interactions, allowing human teams to focus on strategic and empathy-driven responsibilities. The curiosity around GenAI will eventually lead enterprises to explore Agentic AI which is evident in the Deloitte’s prediction- “25% of enterprises using GenAI are forecast to deploy AI agents in 2025, growing to 50% by 2027.”
Agentic AI can be broadly categorized under:
Tool-Based Agents
AI systems designed to perform specific tasks by utilizing various software tools or applications. These agents can automate processes, assist users in completing particular tasks, or enhance productivity by integrating with existing software applications (e.g., CRM systems, project management tools). Major companies like Anthropic have already released these, with OpenAI planning a release in January 2025.
Simulation Agents
AI systems that model real-world processes or environments to study their behavior under various conditions. Simulation agents create representations of real-world systems (e.g., traffic flow, weather patterns) to analyze their dynamics and interactions. They can also mimic human behavior and replicate people’s values and preferences.
With the rise of Agentic AI across diverse organizational workflows especially in the form of Agentic mesh – a network of specialized AI agents- Gartner also warns about the repercussions. Enterprises must invest in the skills, practices and technologies to deliver trustworthy AI agents. Ensuring a supply of high-quality data is essential as Agentic AI will make decisions based on its analysis of your organization’s data.
Multimodal AI
It’s been more than two years since the launch of first public GenAI (Generative AI) solution, Chat GPT was launched. ChatGPT and other subsequent publicly released GenAI solution entertains interaction in single modality which is text-based as they are trained to interpret. 2025 will see the rise in multimodal AIs offering outputs in diverse range of modalities like text to graphs, voice clip to images and so on. We will witness the blending of multimodal AI with supervised learning. Data from inputs of diverse modalities will get annotated enabling AI to detect sarcasm in text or emotions in video.
The possibilities that multimodal AI exhibits are endless. In a healthcare startup, it can shift through years of sleep related data and combinedly analyze MRIs of heart to track for signals of sleep apnea. For the marketing team it can scan through survey forms, heatmaps, and customer browsing history to uncover insights related to customer preference.
Composite AI
As enterprises look to find the right kind of AI solution to their business needs it becomes evident that a singular AI technique can’t handle complex real-world problems. To address this the concept of “composite AI” is gaining popularity and will become a frontrunner of AI trends of 2025. As Gartner defines it as “the combination and application (fusion) of different AI technologies to improve learning efficiency and expand the level of knowledge representation.”
Composite AI merges different AI technologies like machine learning, natural language processing, and knowledge graphs to solve complex business problems breaking them down into multiple sub-problems. They can be broadly categorized under two types:
Composite AI Framework
An example of framework is, Microsoft’s Semantic Kernel, a lightweight SDK that merges LLM prompts with traditional programming languages for application development. It provides a mechanism for the kernel’s orchestration function to accomplish its goals via the skillmemory connector based on user requirements. This facilitates integration into Microsoft’s Copilot offering and enables rapid development of enterprise-specific AI applications.
Composite AI Orchestration
HuggingGPT is an example of orchestration. It automatically selects and a model with ChatGPT and executes from a variety models available on HuggingFace. Multimodal processing of text, image, and voice data takes place by combining multiple models.
AGI (Artificial General Intelligence)
Sam Altman, CEO of OpenAI started 2025 with a bold declaration: OpenAI has figured out how to create artificial general intelligence (AGI), positioning it as one of the most significant technological trends of the year. AGI, characterized by its ability to understand, learn, and apply knowledge across various domains like a human, is expected to revolutionize industries and everyday life.
AGI (also referred to as strong AI or deep AI) is based on the theory of mind AI framework. Fundamentally, the theory of mind-level AI deals with training machines to learn human behavior and understand the fundamental aspects of consciousness. With such a strong AI foundation, AGI can plan, learn cognitive abilities, make judgments, handle uncertain situations, and integrate prior knowledge in decision making or improve accuracy. Existing artificial intelligence capabilities are referred to as narrow AI when compared with artificial general intelligence. AGI is considered to be strong artificial intelligence (AI). Strong AI contrasts with weak or narrow AI, which is the application of artificial intelligence to specific tasks or problems. AGI facilitates machines to perform innovative, imaginative, and creative tasks.
AGI promises to enhance collaboration between humans and machines. AGI should theoretically be able to perform any task that a human can and exhibit a range of intelligence in different areas without human intervention. Its performance should be as good as or better than humans at solving problems in most areas. In industries like manufacturing, AGI systems could work alongside human operators to optimize production processes and improve efficiency.
Edge AI or Edge Intelligence
As a key driver that boosts AI development, big data has recently gone through a radical shift of data sources from mega-scale cloud data centers to increasingly widespread end devices, such as mobile, edge, and IoT devices. Traditionally, big data, such as online shopping records, social media content, and business informatics, were mainly born and stored at mega-scale data centers. However, with the emergence of mobile computing and IoT, the trend is reversing now.
IDC forecasts that, by 2025, 80 billion IoT devices and sensors will be online. Moving a tremendous amount of collected data across the wide-area network (WAN) poses serious challenges to network capacity and the computing power of cloud computing infrastructures. Many new types of applications have challenging delay requirements that the cloud would have difficulty meeting consistently (e.g., cooperative autonomous driving).
Edge Computing is a paradigm to push cloud services from the network core to the network edges. The goal of Edge Computing is to host computation tasks as close as possible to the data sources and end-users. The combination of Edge Computing and AI has given rise to a new research area named “Edge Intelligence” or “Edge ML”. Edge Intelligence makes use of the widespread edge resources to power AI applications without entirely relying on the cloud.
The Gartner Hype Cycles names Edge Intelligence as an emerging technology that will reach a plateau of productivity in the following 5 to 10 years. Multiple major enterprises and technology leaders, including Google, Microsoft, IBM, and Intel, demonstrated the advantages of edge computing in bridging the last mile of AI. This form of AI utilizes the processing power of the device’s own hardware, such as CPUs, GPUs, or specialized chips like neural processing units (NPUs), to run AI algorithms locally, also known as on-device AI.
Generative Virtual Playgrounds
In October 2024, AI startups Decart and Etched unveiled an unofficial Minecraft hack where each frame of the game is generated dynamically as players engage. If 2023 marked the era of generative images and 2024 highlighted generative video, we are now stepping into the realm of generative virtual worlds.
AI has the potential to significantly cut development time and costs by automating various world-building processes:
- Procedural Generation: AI algorithms can create expansive and varied virtual environments, from vast cities to alien landscapes. By establishing rules and constraints, AI can unlock limitless creative possibilities.
- AI-Generated Narratives: AI can craft adaptive, dynamic storylines that respond to players’ decisions, enabling highly personalized and immersive storytelling.
- AI-Driven Character Development: AI can design lifelike characters with distinct personalities, histories, and behaviors, enriching social interactions within the metaverse.
The convergence of AI with virtual and augmented reality is making it increasingly challenging to distinguish between real and virtual environments. Generative AI is poised to deliver deeply immersive virtual experiences.
- Real-Time Content Creation: AI can generate content on the fly as users explore the metaverse, ensuring a continuous flow of novel and engaging experiences.
- AI-Driven Personalization: By analyzing user preferences and behavior, AI can customize virtual worlds to create unique, tailored experiences for each individual.
- AI-Enhanced Social Interactions: AI can support seamless and meaningful social connections among users, fostering the development of communities and friendships.
Conclusion
AI is rapidly evolving, with new use cases emerging daily across various industries. While AI drives significant advancements, efforts are also underway to enhance its intelligence, safety, and ethical alignment. McKinsey estimates that AI-driven productivity improvements could contribute $340 billion annually to the banking sector alone if fully implemented.
As AI adoption grows, ethical considerations and sustainability are becoming increasingly important. Companies like Anthropic, founded by former OpenAI members, emphasize ethical AI development. Over 100 universal basic income (UBI) pilots are underway in the U.S., with prominent figures like Geoffrey Hinton, the godfather of AI, advocating for such initiatives. Despite the substantial computational resources AI requires—which have environmental implications—AI has also enabled significant progress toward sustainability goals. Especially, as PwC reports that using AI for sustainability applications could reduce global greenhouse gas emissions by as much as 4% by 2030.
