AI is only as trustworthy as the openness of the internet

World leaders will gather in New York this week for the 81st UN General Assembly, which will gather under the theme “Restoring Trust, Managing Transformation.” While the rise of artificial intelligence is bound to be at the top of their agenda, rebuilding trust in technology also requires that we protect the open internet, which has made much of today’s technological progress possible in the first place.

There is no denying that trust in technology is declining. Increasingly, people feel that technology is using them rather than the other way round. Their personal data is used to train large language models and their communities bear the costs of grid-intensive data centers, while every day brings a new barrage of online scams and fraud.

Meanwhile, the world’s most powerful tech companies are making enormously consequential decisions, often with little input from technical experts and the wider public. Frontier AI, for example, is being shaped by a small number of governments and companies rushing to consolidate their dominance. Researchers and civil society groups might be invited to comment but it remains woefully unclear how, or even whether, their input is considered.

Protecting the public interest does not have to come at the expense of innovation. The internet is proof of that
Sally Wentworth

Yet protecting the public interest does not have to come at the expense of innovation. The internet itself is proof of that; its multistakeholder model shows how a profound technological shift can be managed in a way that serves the public interest while creating enormous economic value.

The internet was not built by any single government or company; it grew out of cooperation among governments, businesses and civil society. That cooperation helped create and sustain the building blocks of the global internet as we know it today: open and interoperable technical standards, secure systems and resilient infrastructure.

The World Wide Web is a prime example. Developed at the European Organization for Nuclear Research, also known as CERN, it began as a way for scientists to share information across institutions. In 1993, CERN made web software freely available, relinquishing its intellectual property rights so that anyone could use, modify and distribute it. That decision helped turn a tool built for researchers into a force that continues to transform economies and societies.

There is a useful lesson here for the AI era: Transformative technologies spread and develop faster when the infrastructure they rely on is open and people around the world can build on it. Many of the technologies that facilitated the rise of AI, including those originally developed to make the internet more accessible and participatory, owe much to this tradition of public-minded collaboration across borders.

That is why the Internet Society has spent decades working with stakeholders at every level, from local communities to the UN, to help ensure that the internet we all rely on remains open and interoperable. While this approach has its flaws — the industry moves fast, while the slower process of reaching consensus can struggle to keep up — the alternative of ceding decision-making to a handful of powerful governments and companies is far worse.

Rebuilding trust in technology requires more than making individual systems safer or more reliable
Sally Wentworth

Regrettably, that appears to be where we are headed. As industry consolidation puts control of the internet in fewer hands, geopolitical fragmentation is driving governments to carve it up. These changes in turn threaten every technology built on top of it, particularly AI.

The consequences are already visible. “Walled gardens” (closed tech ecosystems where a single company tightly controls hardware, software, services or data) and national intranets restrict what people can access, turning a global network into a patchwork of confined spaces, and government-mandated internet shutdowns can cut people off from economic and civic life for days at a time — or even longer. Many communities still cannot even get online at all because of inadequate infrastructure.

Rebuilding trust in technology requires more than making individual systems safer or more reliable. We also need to protect the foundation on which these technologies depend and ensure that its future is not decided by corporate and government interests.

An internet that remains open and accessible can support a thriving digital economy in which ideas and innovation travel freely across borders, giving people everywhere the ability to develop new technologies that serve the public good.

The birth of the internet showed what can happen when we get those conditions right. Its openness has given generations of innovators the freedom to experiment and create things its architects could never have imagined. As AI gathers pace, we should remember that this openness is a prerequisite for innovation, not an impediment to it.