No one has a crystal ball to predict the future, but the future of AI in 2025 looks bright.
getty
The past 12 months have seen significant advances across AI. These include everything from building large-scale supercomputers for continuous AI training, to large-scale language models to learn how to use computers, to extending AI to internet searches and real-time political polling. Includes things.
For this prediction perspective, five experts were asked to share their top three AI predictions.
Based on insights gleaned from this small group of AI experts, 2025 could be an even more revolutionary year for the field than 2024, with some surprising predictions and trends likely to unfold by the end of the new year. There is.
AI agents mature
Nearly every expert cited the rise of AI agents, computer programs that can think, learn, and act autonomously, as the most important trend to watch next year.
Pascal Bornet is an award-winning expert, keynote speaker, and author. His latest book, IRREPLACEABLE: The Art of Standout in the Age of Artificial Intelligence, distills his research into practical strategies. He says that by 2025, the limits of AI will be pushed beyond our expectations.
“Agent AI is redefining autonomy. These systems don’t just respond, they proactively anticipate our needs, act autonomously, solve problems before we even know it, and take care of work and work. They become indispensable collaborators in both our lives,” Bornet shared via email.
Ben Torben-Nielsen, Ph.D., MBA, holds two machine learning patents and is an internationally recognized enterprise AI consultant. He agrees that AI agents have transformative potential, but when combined with cheaper miniature language models, they add a twist.
“To solve business problems, managers need to stitch together many tools. Beyond coding, generative AI is rarely integrated into other workflows. (combined with large-scale language models) provides a solution by letting AI actually use these tools, and this is where the real potential lies,” Torben Nielsen explained in an email. did.
Meanwhile, Eduardo Ordax, who has more than 10 years of experience in the field with a focus on big data, analytics, and artificial intelligence under internationally recognized leaders in the field, believes that AI agents can build language models at scale. We predict that we have the potential to triple the specific benefits realized from
“While current LLMs can increase productivity by 20-30%, AI agents have the potential to double or triple these benefits in certain use cases.More autonomous AI systems , which could lead to greater adoption across the industry by streamlining operations and reducing costs,” Ordax said in a text message.
AI as a competitive advantage
Dr. Ahmed Banafa, a technology expert and professor of engineering at San Jose State University, shared via email that generative AI will be safely and securely integrated into enterprise software.
“Businesses will leverage these models to create content, interact with customers, and optimize workflows. This change is being driven by advances in AI capabilities and their increasing cost-effectiveness, and AI is being used in a variety of ways. It has become essential to remain competitive in the field,” Banafa explained.
Julia McCoy, who started her own hybrid AI technology consulting firm, First Movers, writes that her AI clients are making incredible profits in text messaging, giving them a distinct advantage over their competitors.
“Companies that successfully implement AI-driven processes will be able to reduce their workforce by more than 95% while maintaining or increasing production volumes. “This means a fundamental restructuring of how business is run and where value is created,” she wrote.
Torben-Nielsen was even more dire and direct for AI layoffs.
“Ignoring AI can put a company out of business. We saw a similar wave of business closures when companies didn’t adopt e-commerce, and it will happen with AI as well. It will quickly become extinct.” As PwC wrote, “Your AI strategy will get you ahead, or you’ll have a hard time catching up.” ”’ he added.
AI costs are likely to increase significantly
He was also the only expert to predict a sharp increase in prices for companies looking to integrate AI in 2025.
“The cost of operating AI is becoming much more expensive. First, the high-end models will be more expensive. OpenAI is experimenting with this, and other companies will likely follow. Second, per token (the text-based components that LLM uses for queries and basic pricing) are plummeting, while other models are likely to experience a sharp increase in token usage and higher operating costs due to increased adoption. There are many companies. You are likely to have underestimated your AI operating budget for 2025 and will find yourself facing unexpectedly high bills,” Torben Nielsen warned.
AI and robot integration
Bornet and Banafa said AI is likely to move away from desktops and mobile devices and toward mobile robots that interact more frequently in the real world.
“Robots will no longer be futuristic, they will become everyday. Wearables and smart devices will blend seamlessly into our environment, and AI will become tactile and tangible. These advances will tackle ever more complex challenges, from global logistics to personal accessibility, and lay the foundations for a better, more just world,” Bornet wrote.
“AI systems that can process and integrate different data types, such as text, images, audio, and video, can significantly improve their understanding of context. , applications in fields such as entertainment will be revolutionized. This progress will be facilitated by advances in computing power and the increasing availability of large-scale, multimodal datasets,” Banafa explained. I did.
Adoption of stricter AI regulations
Banafa was the only expert to discuss the topic of AI regulation, specifically mentioning the possibility of future crackdowns by regulators.
“Global AI regulation will become more stringent, emphasizing transparency, ethical use, and accountability. The need to address bias, data privacy, and potential for abuse will encourage companies to innovate responsibly. “We must prioritize compliance in parallel. This may slow some unregulated progress, but will ultimately ensure safer and more reliable AI applications.” I concluded.
AI integration and shakeout
Meanwhile, Ordax mentioned the possibility of business model changes and possible mergers between generative AI providers to remain competitive.
“Currently, training and running an LLM requires significant investment in expertise, data collection, and computing. To increase profits, many AI labs are moving away from pure-play LLM providers to direct-to-end-user services. “This transformation will likely lead to market consolidation, with a few large AI research companies dominating the space,” he concluded.
