AI Talks

Applications and Emerging Technologies

2024 witnessed significant advancements across various areas of AI, from sector-wide transformations driven by multimodal applications to the rise of AI agents, the evolution of language models, and the increasing prevalence of on-device intelligence and AI-powered productivity tools.

Sector-Wide Transformation

AI’s practical applications became increasingly evident across various sectors in 2024, impacting everything from drug discovery and financial workflows to robotics and content creation.

• Healthcare: AI made significant strides in healthcare, particularly in drug discovery and disease prediction. DeepMind expressed optimism that AI-generated prescription drugs were nearing clinical trials, while Boston University researchers developed an AI tool to predict Alzheimer’s risk with approximately 80% accuracy through speech patterns.
• Financial Services: The financial services sector saw significant AI integration such as Klarna’s use of AI agents to improve customer support, Morgan Stanley’s implementation of AskResearchGPT for enhanced access to research insights, and Microsoft’s launch of Copilot for Finance to streamline financial workflows.
• Manufacturing and Robotics: Significant research advanced the capabilities of AI in manufacturing and robotics. For example, the Chinese company LimX Dynamics used generative AI to accelerate its research and development cycle, shortening the path to practical humanoid robots. Boston Dynamics and the Toyota Research Institute (TRI) planned to integrate large behavior models (LBMs), similar to large language models (LLMs), into the Atlas humanoid robot to move it closer to achieving general-purpose capabilities. Researchers at the University of Virginia pioneered an AI-driven system using Multi-Agent Reinforcement Learning (MARL) to optimize manufacturing processes, with broad applicability across industries.
• Media and Entertainment: AI continued to transform media and entertainment, with personalized content and generative AI for content creation as key trends. This was demonstrated by advancements in various AI-powered content creation tools, including video generation tools like OpenAI’s Sora and Runway ML’s Gen-3 Alpha (generating high-quality video from text prompts) and audio/podcast production tools like Google’s NotebookLM (generating audio overview from various document types) and ElevenLabs’ GEnFM (creating AI-powered multi-speaker podcasts).

Small vs. Large Language Models (SLMs vs. LLMs)

While large language models continued to advance in capabilities, with notable releases like Google’s Gemini 2.0 and OpenAI’s GPT-4o, major players also invested in developing lighter, more efficient models better suited for deployment on edge devices and in resource-constrained environments due to their improved size-performance trade-off. These smaller models included Microsoft’s Phi-4, OpenAI’s o1-mini, Meta’s Llama 3.2 and Google’s Gemma.

The Rise of AI Agents

In 2024, AI agents advanced significantly, with new development frameworks like LangGraph enabling more sophisticated and adaptable agents and increasing autonomy allowing for less human oversight. The growing trend of independent and proactive “agentic AI” gained significant traction, with a majority of surveyed companies planning to integrate AI agents into their business operations for greater efficiency and cost savings.

On-Device Intelligence

The trend of On-Device Intelligence significantly enhanced the user experience across smartphones and PCs. By enabling complex AI tasks directly on devices through dedicated AI hardware, manufacturers delivered faster processing, improved security, and greater responsiveness. This trend, already evident in late 2023 with devices like the Google Pixel 8 series, continued with new device launches from major players like Samsung (Galaxy S24), Apple (latest iPhones and Macs), and Microsoft (Copilot+ PCs).

AI Productivity Tools

 AI-powered productivity tools continued to streamline workplace workflows. For example, Microsoft introduced agent-based automation with Microsoft 365 Copilot Studio, empowering users to create and manage autonomous Copilot agents that execute work on their behalf, while Google enhanced its AI-powered productivity tools in Workspace by integrating Gemini, its most capable AI model, directly into Workspace applications, streamlining tasks such as writing emails, summarizing documents, and information retrieval.

Infrastructure, Access, and Data

Significant developments in infrastructure, access to AI resources, and the crucial role of data shaped the AI landscape throughout 2024.

Technological Infrastructure

The AI accelerator market remained highly competitive in 2024 as AI infrastructure continued to evolve, with Nvidia maintaining its leadership position through ongoing chip design refinements and ecosystem expansion. Groq distinguished itself by its focus on high-performance, low-latency AI inference, further expanding access to its specialized AI hardware via GroqCloud. This evolution also saw a growing focus on sustainable energy solutions, such as Small Modular Reactors (SMRs) and  larger nuclear reactors, to address AI’s increasing energy demands. Optical computing, which aims to revolutionize data transmission using light, achieved a key milestone with Intel’s demonstration of its first fully integrated optical I/O chiplet.

Open Source vs. Closed Models

The tension between accessibility and control continued to shape the AI landscape. While Meta’s CEO, Mark Zuckerberg championed accessibility  by publicly supporting open-source AI, the interpretation of “open-source” in this context was debated in light of the newly established Open Source AI Definition. Major players like OpenAI maintained closed, proprietary models, prioritizing control over open access. Google demonstrated a more nuanced approach with the release of Gemma, a family of lightweight, open models, while maintaining its more powerful and capable models as proprietary offerings.

Data as a Key Resource

In 2024, the escalating demand for data to fuel AI development reached a critical juncture. While data remained the lifeblood of AI, driving significant data acquisition deals between tech companies and content owners, the specter of data scarcity loomed large. This scramble for existing data resources intensified concerns about data privacy and ethical data collection practices, as companies sought training datasets. Simultaneously, the growing recognition of impending data limitations, particularly the potential exhaustion of high-quality public domain data, accelerated the development and adoption of synthetic data for training and evaluating AI agents. This shift towards synthetic data represented a strategic move to mitigate the challenges posed by both data scarcity and the ethical and legal complexities surrounding the acquisition and use of real-world data.

Societal, Ethical and Global Dynamics

The year brought intensified focus on the societal and ethical implications of AI’s growing influence. Discussions surrounding job displacement, misinformation, bias, safety, and environmental impact gained prominence, driving the need for responsible AI development and effective governance frameworks.

Economic and Workforce Implications

 The year saw intense discussions about AI’s economic impact. A study commissioned by Microsoft suggested GenAI could significantly boost productivity and economic growth. Companies like Klarna showcased this potential, with AI agents handling tasks equivalent to hundreds of human workers. However, concerns about AI-driven job displacement for both white-collar and entry-level workers became a pressing issue. These concerns fueled discussions about workforce retraining and potential solutions to widespread job disruption, such as universal basic income.

Regulatory Landscape and Ethical Considerations

The European Union’s landmark AI Act emerged as a global benchmark for responsible AI development. This legislation addressed critical concerns around data privacy, algorithmic bias, and potential misuse, setting a precedent for international AI governance. Key ethical areas of focus included:

• Potential job displacement (with some estimates suggesting AI could impact up to three million jobs in the UK, particularly in administrative and customer service roles)
• The role of AI in spreading misinformation during election periods, particularly through deepfakes and manipulated audio/video
• Algorithmic bias in critical decision-making processes (e.g., in mortgage underwriting or police report generation)
• The environmental impact of AI’s massive computational requirements (e.g., the carbon footprint of AI data centers)
• Military and defense applications (the exploration of AI for military purposes, including battle planning and AI-powered weapons systems, without a comprehensive global governance framework for the military use of AI)

Geopolitical Competition

Global competition for AI dominance intensified during the year. The U.S. administration implemented further restrictions on China’s access to advanced semiconductor technology, impacting the development and deployment of cutting-edge AI systems in this country. This “chip war” further complicated the global AI landscape, influencing supply chains, research collaborations, and the overall pace of AI advancement.

Looking Forward

2024 has been a pivotal year for AI, demonstrating its power to reshape industries and unlock new possibilities. As we move forward, focusing on ethical development, mitigating potential risks, and promoting international collaboration will be essential to realizing AI’s full potential for good.

This post was researched and written with the assistance of various AI-based tools.