Technology leaders have warned that organisations, governments and individuals may be underestimating the cybersecurity risks created by artificial intelligence, as increasingly capable systems become more accessible and the cost of deploying advanced computing falls.
The warning emerged during the 2026 CNBC Technology Executive Council Summit in New York, where executives and technology specialists discussed how companies are adapting to the growing role of AI. The event focused on the practical and strategic questions confronting business leaders, including the return on investment from computing infrastructure, cybersecurity, regulation and quantum computing.
Warnings of a worsening threat
OpenAI chief executive Sam Altman said during an appearance at Stanford Graduate School of Business that the developing AI-related cybersecurity crisis remained far from resolved. His comments, reported by Gizmodo and cited in the summit coverage, described a problem that leaders at every level had not yet taken seriously enough.
This cybersecurity…I don't even know a dramatic enough word to use, but the problem coming towards us, I think, is still something that we have not, as organizational leaders or countries or individuals, taken seriously.
Altman’s assessment points to a central difficulty in the debate around artificial intelligence: the technology is advancing while many of the institutions expected to manage its consequences are still working out basic safeguards. The source material does not identify a single incident behind the warning. Instead, it presents the risk as an emerging challenge for businesses, public authorities and individuals.
Campbell Brown, founder and chief executive of Forum AI, offered a similarly cautious outlook at the summit. She predicted a “rocky six months” for the sector, particularly as the public becomes more aware of hacks and leaks that are taking place. Her comments suggest that increased visibility of breaches could alter how people assess the technology, regardless of whether every incident is directly caused by AI.
Executives see the scale changing
Michael Weening, chief executive of Calix, and Shane Eleniak, the company’s chief product officer, went further. They said the problem could become substantially worse and that the potential danger might be greater than current forecasts indicate.
| Speaker | Assessment described at the summit |
|---|---|
| Sam Altman | The approaching cybersecurity problem is not being taken seriously enough. |
| Campbell Brown | The next six months could be difficult as hacks and leaks receive greater public attention. |
| Michael Weening and Shane Eleniak | The threat may worsen considerably and exceed existing expectations. |
Weening said the pace of social change associated with AI was beyond what humanity was prepared for. That is a broader claim than a warning about one software vulnerability: it concerns the speed at which the technology may change the balance between those defending systems and those attempting to compromise them.
Eleniak said large-scale attacks using AI had, so far, largely remained a nation-state issue. He attributed that in part to the expense involved in assembling such operations. The distinction matters because a capability that is currently limited by cost could become more widely available if the underlying tools and hardware become cheaper.
Lower barriers raise questions
He pointed to the availability of open-weight models and more economical graphics processing units as factors in that changing environment. Open-weight systems can be accessed and adapted more freely than tightly controlled models, while GPUs provide much of the computing power used by modern AI applications. The source content ends while describing the significance of those cheaper resources, but the direction of the warning is clear: lower technical or financial barriers could widen access to powerful capabilities.
For organisations, the immediate challenge is not simply deciding whether to use AI. Leaders must also consider how systems are protected, how sensitive information is handled and how quickly defensive measures can respond to new forms of attack. The summit’s wider agenda placed those questions alongside regulation and computing economics, underlining that security is part of the business case for AI rather than a separate technical concern.
The discussions also show why claims about AI need to be tested against evidence. The speakers offered warnings about the pace and possible scale of future attacks, but the source does not provide a forecast, a list of affected organisations or figures quantifying the threat. What it does establish is a shared concern among senior figures that the risks are evolving faster than public understanding and institutional readiness.
- AI security risks were described as an expanding challenge for companies, governments and individuals.
- Executives warned that the next six months could bring greater public awareness of hacks and leaks.
- Cheaper computing and more accessible models could reduce the cost of mounting sophisticated attacks.
