Slowing Down AI ‘Unrealistic’: Palo Alto CEO Nikesh Arora

Nikesh Arora Calls AI Slowdown Unrealistic

Palo Alto Networks CEO Nikesh Arora has pushed back against calls for a coordinated slowdown in artificial intelligence development, describing the idea as unrealistic.

Speaking on CNBC's “The Tech Download” podcast, Arora said the difficulty is that not every company developing AI will necessarily agree to slow down at the same time.

His comments come as technology companies, researchers and policymakers debate how quickly frontier AI systems should be developed and deployed.

Why Does Arora Think an AI Slowdown Is Difficult?

Arora's argument centres on competition.

If some AI companies voluntarily reduce the pace of development while competitors continue investing heavily in increasingly capable models, the companies that slow down could potentially fall behind.

He therefore questioned whether a voluntary, industry-wide slowdown could work when companies have different approaches to AI development and safety.

Debate Over AI Safety Is Growing

Arora's comments come during a period of increased attention to AI safety.

Several AI leaders and researchers have recently discussed whether the pace of AI development should be moderated and whether stronger safeguards or independent assessments are needed.

A separate recent analysis noted that Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have supported discussions around slowing or reassessing the pace of frontier AI development.

The debate therefore involves competing considerations: technological progress, commercial competition, security and potential risks from increasingly capable AI systems.

Arora on the Risk of AI Wiping Out Humanity

Arora was also asked about concerns that AI could eventually pose an existential threat to humanity.

He described a 10% probability of AI wiping out humanity as highly unlikely, while not saying that the possibility was absolutely zero.

His comments represent his personal assessment rather than an established scientific probability.

US-China Coordination Also Raises Questions

The discussion around slowing AI development extends beyond individual companies.

AI development is taking place in multiple countries, with the United States and China among the major centres of investment and research.

Arora questioned whether countries with competing strategic and technological interests could realistically coordinate on the pace and safety of AI development. This adds another layer of difficulty to proposals for internationally coordinated AI restrictions.

Cybersecurity Becomes More Important as AI Advances

As AI capabilities increase, cybersecurity is becoming closely connected to the development of AI systems.

Palo Alto Networks has been developing AI-based cybersecurity products and services, including its Unit 42 Continuous Frontier AI Defense, announced earlier this week. The service uses advanced AI models to continuously identify vulnerabilities and potential attack paths across applications, APIs and cloud infrastructure.

This reflects a broader industry trend in which AI is being used both to create new capabilities and to address security threats created or accelerated by AI.

AI Development and Regulation Remain in Tension

The current debate is not simply about whether AI should be developed faster or slower.

It also concerns how governments and technology companies can establish safety standards while allowing research and innovation to continue.

Recent proposals discussed by AI researchers have included independent evaluations, model audits, monitoring of computing resources and stronger incident-reporting mechanisms.

Different technology leaders have taken different positions on how such safeguards should be implemented.

What Arora's Comments Mean for the AI Debate

Arora's comments highlight one of the central challenges facing policymakers: a slowdown is difficult to coordinate when AI development is highly competitive and international.

Even if some companies voluntarily adopt slower development schedules or additional safety checks, competitors may continue moving ahead.

At the same time, concerns about AI safety have prompted calls for stronger testing, monitoring and oversight. The debate over the appropriate balance between innovation and safety is therefore continuing.

Key Takeaways

• Palo Alto Networks CEO Nikesh Arora called attempts to slow AI development collectively “unrealistic”.

• He said not every AI company is likely to voluntarily reduce its development pace.

• His comments come amid growing debate over frontier AI safety and regulation.

• Arora described a 10% chance of AI wiping out humanity as highly unlikely, while not ruling out the possibility entirely.

• He also questioned whether the US and China could coordinate effectively on AI safety.

• AI development is increasingly linked to cybersecurity risks and opportunities.

• The debate over AI safety involves companies, governments and researchers with differing approaches.

• Arora's comments are his own views and do not establish a scientific probability or policy conclusion.