What to expect from AI in 2025, according to industry leaders

• 2024 was a big year for artificial intelligence. 2025 could be even bigger.
• Business Insider spoke to over a dozen key figures in the industry about AI's future.
• Here's what they had to say.

If 2024 is the year companies started adopting AI, then 2025 could be the year they start tailoring it to fit their needs.

Some say AI will become so integrated into our lives we won't even notice it's there.

"Like the internet or electricity, AI will become an invisible driver of outcomes, not a selling point," Tom Biegala, cofounder of Bison Ventures, a venture firm focused on frontier technology, told Business Insider by email.

And as companies incorporate the technology into their businesses, they'll likely need to focus more on managing it responsibly.

"In 2025 we expect more enterprise companies will recognize that investing in AI governance is just as important as adopting AI itself," Navrina Singh, founder of Credo AI, an AI governance platform, said.

Business Insider spoke with 13 key figures in tech — from startup founders to investors — for their best guesses on what to expect from AI in 2025.

Investment will continue to soar.

"The AI hype cycle may stabilize, but AI investments will soar," Immad Akhund, the CEO of Mercury, which offers banking services to startups, told BI by email.

He believes the sustained interest in AI comes as companies move from experimenting to using it in real-world areas like customer service, sales, and finance.

"Companies will use AI to boost productivity — especially in back-office tasks and document management — helping small teams scale quickly and operate more efficiently," he said.

Under the Trump Administration, the new leadership at the Federal Trade Commission might foster a more favorable climate for mergers, acquisitions, and IPOs in the AI industry.

"I expect M&A to increase by at least 35% next year," Tomasz Tunguz, founder of Theory Ventures, a venture capital firm, told BI. "The top 10 most active acquirers in the software world are falling off a cliff in terms of activity, which requires meaningfully the IPO market to roar open with a combination of AI and other software companies."

The competition will get fierce.

Don't be surprised if a leading company takes a hit because of AI.

"At least one major, globally recognized company will fail or significantly downsize due to an inability to compete with one or more AI-native startups. Rapid innovation cycles and the horizontal application of AI will render slow movers obsolete," Stefan Weitz, CEO and cofounder of HumanX, a leading AI conference, told BI.

He believes the tech's threat will extend to the global stage, requiring major powers to regulate AI to maintain their competitive edge.

"As we are already seeing with the US and China regulating or blocking core AI technologies, nations or corporations will experience major geopolitical conflicts over AI algorithms and data, with some countries banning or nationalizing key AI technologies to maintain control over economic and political power," he wrote.

That said, the United States and China are already working together to mitigate the existential threat AI poses to humanity. In November, at the Asia-Pacific Economic Cooperation Summit, President Joe Biden and Chinese leader Xi Jinping agreed that humans, not AI, should make decisions regarding the use of nuclear technology.

The lines between humans and AI will not be obvious.

The idea of humans and autonomous agents working together might soon move beyond the realm of science fiction. That means we'll also need to start drafting rules to govern these interactions.

"Synthetic virtual people indistinguishable from real humans will enter the workforce, even if in limited ways, leading to debates about employment rights and creating a push for 'AI citizenship' to define their societal roles and limitations," Weitz said.

Some predict that the distinction between human-created and AI-generated content will also become increasingly unclear.

"Generative media will hit the mainstream in a big way and will be as much talked about as LLMs in 2024," Steve Jang, founder and managing partner of Kindred Ventures, an early-stage venture firm, told BI. "Generative audio and images are getting better due to more advanced models, and we'll start to see adoption spike across both consumer and enterprise."

Specialization. Specialization. Specialization.

Business leaders told BI that next year will be about custom-fitting AI technology to suit specific needs.

"In 2025, the AI hype cycle will give way to the rise of domain-specific, specialized AI and robotics," Biegala said. "Products will be faster and more efficient while delivering immediate, tangible value compared to general-purpose solutions. This shift will mark the beginning of real, transformative economic impact of AI."

The focus on customization also extends to how we search for information online, with chatbots replacing search engines like Google.

"In 2025, search will no longer be synonymous with a single brand; instead, users will turn to multiple platforms for specific types of queries. Some may rely on AI-powered chatbots for conversational answers, others on domain-specific engines for technical or industry-specific expertise, and still others on visual or voice-based tools for multimedia queries," Dominik Mazur, CEO and cofounder of IAsk, an AI search engine, told BI. "This diversification will create a competitive environment where specialized players and niche solutions coexist with larger generalist platforms, leading to greater innovation and choice for users."

Over the past year, AI leaders have been promoting the value of smaller AI models that can address a company's specific needs better than large-scale foundation models. "There's a lot of pressure on making smaller, more efficient models, smarter via data and algorithms, methods, rather than just scaling up due to market forces," Aidan Gomez, the founder and CEO of Cohere, an enterprise AI startup, previously told BI.

The pressure is rising as the value of building models simply based on computing power decreases.

"The days of using a GPU to brute force compute to build models and applications will be in the rearview mirror," Biegala said.

Companies may also use customizable AI tools more, possibly replacing software-as-a-service applications.

"AI tools are tearing down the moat of SaaS applications as tools that can only be bought vs built, prompting enterprises — from Amazon to ambitious startups — to replace expensive SaaS apps that don't quite totally fit the need with lightweight custom-fit solutions integrated into your stack," David Hsu, founder of Retool, a low code platform for developers, told BI.

Regulation takes priority.

With more responsibility comes more risk. Companies are going to start getting serious about regulation.

"I expect to see more voluntary commitments and actions to responsible AI. I think there will be a push to establish guardrails similar to what happened for frontier models, now discussed for AI agents and autonomous AI," Singh said. "Also, I do see a world where we will see the first penalties for noncompliance with AI-specific laws, which will set a global precedent, forcing businesses to prioritize governance or face steep consequences."

Singh, along with others like AI godfather Geoffrey Hinton and OpenAI CEO Sam Altman, have expressed interest in an international body to govern the use of AI. We may "even see Global AI standards emerge, led by coalitions of nations and enterprises to set the baseline for safety, transparency, and accountability in AI systems," she said.

The value of regulation will be paramount next year amid the growing threat of large-scale AI-driven cybersecurity threats.

"AI deepfake technologies will make generating fake identities and documents trivially easy, creating a trust crisis for businesses," Pat Kinsel, the CEO of Proof, a software platform for notarization, told BI. "The ability to distinguish between real and fraudulent identities and secure digital interactions in the AI age will be the key differentiator between resilient businesses and those at risk of costly fraud."

Search advertising will get conversational.

It's typically easy to distinguish sponsored content from relevant search findings in an online search. This year, though, sponsored content may blend seamlessly into chatbot answers.

"Generative AI will disrupt the traditional search advertising model by integrating relevant sponsored content directly into synthesized responses," IAsk's Mazur told BI. "This shift toward context-aware, conversational advertising will create a smoother user experience, as ads become less intrusive and more aligned with users' needs."

The challenge, he added, is that companies will need to maintain transparency and trust as users gain awareness of how their queries inform monetization strategies.

AI will not take your job — yet.

The good news is that business and tech leaders only expect to see AI enhance people's occupations next year, not replace them.

"We'll see efficiency gains in industries that automate repetitive tasks, but humans will still be needed for complex decision-making and creative work. 2025 is the year we really see many using AI as a core part of their job and enabling more productivity," Akhund said.