We’ve seen an unprecedented acceleration of AI changes in 2025, with generative AI, agentic automation, and AI infrastructure at the forefront. Investment surged, major tech companies released new models, and the integration of AI in everyday life deepened dramatically.
So let’s dig into those changes, the public perception of AI, and how institutions should respond to AI changes in 2025.
AI Changes in 2025: A Year of Rapid Breakthroughs
Generative AI reached new heights in recent months, with models like GPT-5 setting a new bar for reasoning, memory, and personalization. Agentic AI platforms from companies like Amazon and Microsoft have enabled multistep workflow automation across enterprises.
At the same time, innovation wasn’t limited to the largest models. Google’s Gemma 3 and other open-source models brought efficient, nimble AI to edge devices. Deep integration has also defined the year, as tools like Microsoft Copilot became part of everyday workflows, offering context-aware support and automation.
These advancements translated into tangible breakthroughs across industries. Healthcare systems needed less data for accurate diagnoses. Classrooms benefited from personalized feedback. Scientific fields like materials science and drug discovery accelerated their progress thanks to AI-driven insights.
Despite the rapid innovation, a new MIT study reveals a significant disconnect between AI hype and business reality. In fact, 95% of enterprise AI pilots are failing to deliver measurable ROI.
Why? In many cases, expectations were unrealistic. Businesses assumed that generic, off-the-shelf AI tools could be dropped into existing systems and instantly unlock massive productivity gains. Instead, they encountered the “verification tax” – the need for employees to double-check AI outputs because the models were often “confidently wrong.” This erased promised efficiency and, in many cases, produced negative ROI.
The problems don’t stop there. Many organizations still lack a cohesive strategy, opting for flashy pilot projects that fizzle out rather than scale. AI tools also struggle to adapt to complex workflows, leaving employees frustrated. On top of this, much of enterprise AI spending goes toward sales and marketing instead of back-office functions like logistics, finance, and operations, where automation could actually deliver stronger returns.
Search vs. GenAI: Intent Still Matters
Even as AI transforms how we find and process information, traditional search remains the most powerful tool for lead generation at this time. As of March 2025, Google retains the overwhelming majority of search market share, with AI platforms like ChatGPT holding a nascent, but rapidly growing, market share of approximately 1%.
This dominance matters because user intent on Google is fundamentally different from intent on AI platforms (though Google’s AI overviews are rapidly blurring these lines). Searchers come with a purpose, actively looking for solutions, products, or programs. That makes Google search a richer and more reliable source of high-quality leads, especially in the education sector, where up to 70% of leads that convert into students still originate from paid search ads.
And while OpenAI’s ChatGPT continues to dominate (consistently holding over 75% of the market share), Google Gemini is emerging as a serious contender. Its rise reinforces the importance of staying plugged into the Google ecosystem, where both search and AI are converging.
For now, the message is clear: We must stay aware and participate in the AI game through meaningful SEO, but for now, Google is still where institutions’ primary advertising efforts and budget should be focused.
AI Changes in 2025: Public Perception and Trust
AI experts are far more optimistic and enthusiastic about AI’s potential than the general public. 56% of experts believe AI will have a positive impact over the next 20 years, but only 17% of the public shares that belief.
Concerns outweigh optimism for many users. People worry about job displacement, environmental costs, and a potential loss of creativity and control. These anxieties stand in sharp contrast to expert optimism about AI’s potential for personal and societal benefits.
This public skepticism is also evident when it comes to AI-generated content. As of mid-2024, only about one-third of consumers reported feeling more positively about generative AI tools compared to the year before. A similar portion said their views hadn’t changed, while 22% said they felt more negatively, especially women.
Comfort levels vary by use case. People are uneasy with AI creating virtual celebrity spokespeople, somewhat more accepting of AI editing product images or generating product descriptions, and most comfortable with AI optimizing ad placement.
Awareness is clearly growing, but widespread trust hasn’t yet arrived.
AI Changes in 2025: The Higher Ed Playbook
So how should higher ed marketers use AI?
First, let's address the elephant in the room: As of today, there's still no reliable way to run a self-service ad campaign for a reasonable cost on platforms like Perplexity, ChatGPT, or Google Gemini.
That means it’s time to get thoughtful. Institutions should use this moment to focus on what really matters:
- Pay Attention to How Users Are Spending Their Time With AI: Stay on top of the trends, channel updates, and how people are interacting with these platforms.
- Stand Out for Who You Are: This means staying transparent with audiences about how you’re using AI as well as sharing with your peers what you’re seeing in the marketplace.
- Position Your Institution as an AI Leader: Whether prospects are concerned about AI or embracing it, your institution’s certificates and programs can show them how to stay ahead of the AI curve.
As AI continues to reshape the way we search, learn, and connect, the institutions that stay grounded in strategy while adapting with agility will be the ones that thrive. Looking for advice on crafting or refining that strategy? Schedule a consultation with VisionPoint to discuss how your institution can stay ahead in the AI era while driving real enrollment results.
