3 dynamics reshaping the internet as it reaches an inflection point

The internet is entering a new phase. Three shifts are reshaping who holds power and how trust is maintained: explosive growth in data generation; declining confidence in digital information and transactions; and the rise of artificial intelligence (AI) systems that increasingly mediate access to services and knowledge.

Interviews conducted by the World Economic Forum and KPMG as part of the Connected Future initiative suggest that many institutions, platforms and users are not yet prepared for the pace of change, particularly as AI begins to search, decide and act on behalf of people and organizations.

Participants described existing rules, incentives and accountability mechanisms struggling to keep up, especially around who can use data, who can verify identity and content, and who is responsible when systems act.

This is an inflection point because several assumptions from the previous era no longer hold. The next phase will be shaped by institutional choices and public-private cooperation, as much as by technological capability.

Our research showed that three key dynamics are defining current attitudes around and use of the internet.

For decades, the internet primarily captured what people typed, searched, posted and purchased. That is changing: it is now capturing far more of the physical world, with increasing granularity.

Data is no longer generated mainly through deliberate actions. Sensors, wearables and connected environments create continuous “ambient” (always-on) data – for example, when wearables collect health data that can influence insurance decisions.

This can unlock breakthroughs in health, mobility and public services. It also makes it harder for people to see what is collected, what is inferred and what is shared, raising urgent questions about consent, proof and cross-border governance.

As the pace of development accelerates, the role of AI is also shifting from tool to participant. AI is already reshaping how software is created, enabling dramatically faster development cycles, sometimes reducing timelines to a fraction of what they once were, while also producing code that can exceed human-generated quality.

The rapidly increasing use of autonomous agents raises a practical question: when an agent makes a high-impact decision, who is accountable?

As organizations deploy AI in higher-stakes settings, adoption is increasingly limited not by technical feasibility, but by confidence that systems are safe, compliant and accountable. Legal ambiguity, contractual risk and internal policy are driving demands for clearer liability frameworks, better auditability and, in some cases, indemnification.

The gap between capability and oversight is becoming more visible in high-stakes domains such as payments and financial services. As Larry Wade, Global Head of Compliance at PayPal Crypto and Corporate Staff units, puts it, “Technology is outpacing regulatory capabilities. Significant risks may materialize before effective regulation is implemented, especially in areas involving data, identity and money.”

Synthetic media is also changing the trust environment. Audio, images and video can now convincingly imitate real people and events at scale, making verification harder and faster-moving manipulation easier. The harms extend beyond individual fraud, affecting market integrity, organizational security and the civic information environment when attribution is unclear.

Beyond technical risk the shift is beginning to erode trust in what exactly is truth. Henry Adjer, advisor on deepfakes, warns of creeping “epistemic nihilism”: where people stop believing they can reliably tell what is true. Resilience therefore requires infrastructure that makes trustworthy participation easier than deception.

Nowhere is this trust challenge more acute than in digital identity, where fragmentation limits interoperability across systems and borders. Jordan Burris, Head of Public Sector Strategy at Socure, observes that “digital identity is broken, and in different ways in different countries.”

In practice, this trust deficit shows up in identity and authorization. Countries and sectors verify identity in different ways, so credentials do not travel easily across borders. This makes it harder to accept credentials, move money and verify who is authorized to act, especially as AI agents take on more commercial tasks. The result is an increasingly complex patchwork of national rules.

Digital identity, provenance and authentication are becoming core requirements for participation in markets and public services. Our interviewees noted that AI impersonation could become a dominant fraud vector. In response, identity assurance, secure delegation (authorizing agents) and verifiable records of origin will increasingly determine who can transact, access services and participate online.

AI is becoming an active decision-maker, curating information, prioritizing content and shaping which products and services people see and choose. Organizations increasingly compete not just for consumer attention, but for algorithmic selection, raising new considerations for fairness, transparency and market integrity.

Less visibly, more internet activity is automated. Scrapers, bots and agents index, extract and interact at scale, raising concerns about security, resilience and the economics of a web built for people but increasingly consumed by machines.

At the same time, the nature of internet activity itself is shifting, with automation reshaping how the web is accessed and used. Stephanie Cohen, Chief Strategy Officer at Cloudflare, highlights that "more than 50% of html page requests are now non-human, as humans send agents to access information on their behalf. We're seeing the emergence of two internets – human and machine."

In response, one likely outcome is a shift towards more governed channels, with verifiable bots and clearer protocols for access, attribution and accountability. This will also shape how organizations invest in machine-readable signals of credibility, such as structured data, provenance markers and verifiable claims.

These shifts will change how people interact online, but the shape of this next internet is not predetermined.

The same technologies that can make online life more frictionless can also concentrate power, deepen fragmentation and make trust more costly. The key governance question is simple: who is responsible, individuals, platforms, markets or shared interoperable infrastructure?

Taken together, these shifts are creating a far more complex and less predictable operating environment. The rise of agentic systems, in particular, is expected to bring new forms of uncertainty, including emergent behaviours, opaque interactions and interconnected risk surfaces that regulators and enterprises are only beginning to understand.

Against this backdrop, all three dynamics converge on one question: can the next internet be both powerful and trustworthy? More data, more concern and greater AI intermediation can produce either a virtuous cycle – where trusted agents create value under robust governance – or a vicious one, as opaque systems exploit unregulated data flows while users disengage.

These choices are arriving faster than most participants realize. At this inflection point, the future of the internet will be determined not by capability alone, but by the decisions made now on trust, accountability and governance.