In 2024, artificial intelligence (AI) business continued to surge, with generative AI attracting enormous investments and stirring up huge expectations. Several key moments marked the entire year: the European Union adopted the AI Act, the world’s first comprehensive AI regulation. Open AI, one of the most valuable AI companies, introduced new product offerings. Finally, the world witnessed a rise of AI disinformation attempting to influence political elections.
These developments have left many pondering what 2025 has in store for AI and related technologies. Are we going to witness continued hype and maybe even see more OpenAI-style companies emerging in the market, or will the bubble burst?
7 Predictions for AI in 2025
- AI may not deliver a meaningful ROI yet.
- The generative AI bubble will likely burst.
- Skepticism around AI will lead to further regulation.
- Green and responsible AI will grow in response to environmental/ethical concerns.
- AI and machine learning will fuel scientific inquiry and discovery.
- Multimodal models and text-to-video should improve.
- AutoML will continue to democratize the field.
Problems With ROI Will Persist
Last year, some of us predicted that the generative AI hype would go away already in 2024. And yet, a new influx of money pumped into AI companies without any proof of ROI transformed the dying hype into an even bigger bubble. Even so, in some cases, the technology was not as scalable as investors expected.
Some AI labs are already hinting that scaling the current technology further is not practical, so it is fair to expect that returns from AI capabilities will keep diminishing. Tom Siebel from C3.ai argues that the market is vastly overvaluing the technology, saying, “Absolutely there’s a bubble.”
Moreover, a rising movement of people from diverse professions, including writers, artists, computer scientists, engineers and philosophers, found common ground against generative AI. Skeptics claim that the technology has been forced onto people by billionaires and their organizations and poses irreconcilable problems. Both the increasing regulation to control the negative impacts of technology and public concerns over the perceived dangers of AI should also play into taming enthusiasm around generative AI.
Concerns around the possible negative impact of generative AI should bring such terms as AI safety, responsible AI, and green AI to the forefront of public and technical discussions this year. Servers that support increasingly complex AI models put a strain on the environment. Furthermore, many risks flow from a lack of transparency in how AI is trained and functions — the mysterious black box problem. AI companies must answer these concerns to earn and keep public trust in the benefits of this technology.
Exciting Developments Are Around the Corner
Although we see strong market signals indicating that the generative AI bubble will most likely burst in 2025, many positive developments in this technological area are also on the way. First, we can expect more AI contributions to scientific discovery and academic research in general, first and foremost in the field of data analytics.
ML models already allow researchers who work with large quantitative data sets and experimentation methods to analyze information faster. Furthermore, in the coming years, AI should also help scientists who work with simulations. AI can be used for replicating real-life or hypothetical situations in different scientific fields, from medical research to security and war studies.
Second, multimodal models, especially text-to-video models, should improve significantly in terms of the length of videos, their quality and their obedience to the laws of physics. Currently, AI solutions struggle to create videos longer than a few seconds, especially when requiring high resolution and realistic features. On the other hand, advancements in this AI field should bring heightened scrutiny from regulatory institutions and governments since the technology can be easily used for all sorts of malicious activities, from misinformation and hate crimes to fraud.
Finally, we are hopeful for advancements in automated machine learning (AutoML), which opens this field for those without specialized ML expertise by delegating ML development to algorithms. This means that smaller businesses and organizations that lack financial resources and skilled personnel will have the possibility to implement ML solutions in-house, be it simple business data analytics or automating specific internal processes, such as maintenance.
As such, AutoML is a crucial step toward democratizing machine learning and AI. With more experts in diverse fields being able to create ML tools tailored to their needs, AI and ML adoption in business can accelerate and we might finally see a growing ROI from adopting such solutions.
Don’t Let AI Crowd Out Other Promising Tech
To conclude, AI still offers many fascinating advancements to follow in the near future. Will it be the next big industrial revolution? Time will tell. We shouldn’t forget to look beyond AI to other emerging technologies that might improve human life, however. For example, the progress in decentralization technologies, from the good old blockchain to edge computing, might be no less fascinating to follow.
We believe that the upcoming few years might see paradigm changes in the social systems, such as emerging new technologies that facilitate a decentralized way of life and community-building, for example, decentralized social networks, local currencies, and locally-developed AI, among others. And, as humanity leaps forward with increasingly powerful technological developments, we should remind ourselves that beyond ROI and pure fascination, there is the most important thing we should all strive for — a better way of life.
