SAS, in an official note, wanted to analyze the state of the art of artificial intelligence and the role it will play in the technology sector during 2025. SAS entrusts forecasts and analyzes of sector trends for next year to its executives .
Faster model training reduces the environmental impact of AI
Bryan Harris, Chief Technology Officer of SASunderlines theimportance of speed and algorithmic efficiency to reduce the consumption of cloud resources. “Algorithmic speed and efficiency cannot be ignored as key levers for reducing cloud resource consumption. While AI, known for its high energy requirements, will continue to drive the adoption of sustainable energy sources, such as nuclear, demand for more energy-efficient models is expected to increase. Just as the appliance and automotive industries have made significant progress in terms of energy efficiency, it is essential that AI models also follow this trend.”
AI attacks threaten our way of life
Steven Tiell, Global Head of AI Governance Advisory at SAS, highlights the potential threats of AI on society. “AI’s ability to personalize and operate at scale is changing the way we interact with information, contributing to the spread of misinformation and the manipulation of social norms. AI-based cyber attacks can target individuals, groups or institutions, putting our lifestyles at risk. In this context, democratic societies and their governments have a crucial interest in safeguarding civil debate, elections and cultural norms. To address this challenge, business leaders must take an active role in the debate on the ethical use of AI, promoting corporate values and establishing specific principles, policies, standards and controls related to AI.”
“Data Dumpsters” fuel the AI gap
Marinela Profi, Global GenAI/AI Market Strategy Lead at SAShighlights the importance of quality data. “2025 will show that some organizations are thriving with generative AI, outpacing the competition, creating specialized customer experiences, and launching innovative products more quickly. However, other organizations are falling behind in the race to generative AI. In fact, they are abandoning the series of projects launched in 2023 having overlooked a crucial reality: AI needs quality data. Poor data undermines AI performance, and organizations must have the courage to take a step back to address their data challenges.”
Generative AI: from hype to reality
Jared Peterson, Senior Vice President, Platform Engineering at SASinvites you to focus on concrete value of generative AI. “Generative AI will continue to generate interest, but we have come to a time when it is important to separate the hype and focus on creating real value for businesses. This involves simplifying our approaches, rules and models, complementing them with targeted use of large language models (LLMs) and specialized language models (SLMs).”
Cloud providers and AI users will share environmental responsibility
Jerry Williams, Chief Environmental Officer of SAStalk about shared responsibility in sustainability. “The race to adopt AI is leading to the creation of inefficient models that consume enormous amounts of cloud resources and contribute to a greater environmental impact. It’s not just up to hardware vendors and large cloud providers to reduce environmental impact; it is a shared responsibility with AI users who manage data and workloads. Greater efficiency in AI model development, enabled by cloud-optimized AI and data platforms, will help reduce unnecessary duplication and waste, while minimizing energy consumption.”
The leaders of tomorrow are built today thanks to AI
Jay Upchurch, Chief Information Officer of SAShe predicts a future dominated by AI-enabled organizations. “Fully AI-enabled organizations will be the ones that win the IT battles of 2025. As generative AI evolves from a ‘shiny new toy’ to an established form of AI, companies will begin to operationalize all variants of AI to automate routine tasks, thus freeing employees for higher-value activities. These automations will allow you to make decisions more quickly, identify opportunities more promptly and stimulate innovation compared to competitors. In summary: these organizations will be the real winners.”
LLMs become a commodity and specialize
Udo Sglavo, Vice President, Applied AI & Modeling, R&D at SASunderlines the future commoditization of LLMs. “In 2025, large language models (LLMs) will become a commodity, resulting in the collapse of pricing models for AI as core functionality becomes available for free. The real value will shift towards specialized services and industry-specific applications, developed on this foundation. At the same time, the rise of open-source LLMs will challenge the dominance of some providerser key, fostering a more decentralized AI landscape, where customization and integration will become key differentiators.”
The acceleration of AI and the Cloud will trigger a major rationalization of IT
Stu Bradley, Senior Vice President, Risk, Fraud and Compliance Solutions at SASpredicts a significant simplification of IT infrastructures. “Companies have long operated with isolated systems, each dedicated to a different function or customer segment. IT teams are under pressure from cumbersome integrations, unable to provide the agility their businesses need. A major IT rationalization is looming, where business leaders will use the cloud to simplify IT infrastructures and vendor relationships, gaining critical speed and reducing costs. Those who modernize to a cloud-native platform, powered by AI and capable of supporting multiple functions, will derive the most value. They will be able to acquire integrated and democratized decision-making and data management capabilities, covering the entire lifecycle of the customer and the company as a whole.”
Generative AI gets personal (and more advanced) for marketers
Jennifer Chase, Chief Marketing Officer of SAS, highlights the eevolution of AI in marketing. “In 2025, marketers will rapidly transition from simpler generative AI applications focused on productivity and content creation to more advanced AI capabilities that deliver competitive advantages and revenue growth. Going beyond large language models (LLMs), marketers will adopt generative AI tools such as synthetic data and digital twins, as well as established AI technologies such as machine and deep learning, to deliver personalized experiences and effective campaigns, while respecting the customer privacy.”
Image created with DALL-E
