Dario Amodei, CEO of AI company Anthropic, has proposed a three-stage plan, arguing for slowing the pace of AI model performance improvements to secure time for implementing safety guardrails.
Imagine you are a novice driver who just earned your license, and you are behind the wheel of a high-performance sports car driving at 200 km/h. The car is magnificent and fast, but you haven’t verified whether the brakes work properly or if there are sufficient safety systems to handle such speeds. What if the global race in AI (Artificial Intelligence, computer systems that mimic human intelligence) technology looks exactly like this right now?
Recently, Dario Amodei, CEO of the AI company Anthropic, made a surprising proposal to the AI industry: “slow down” the pace of AI model performance improvements.
Why is this important?
Until now, the AI industry has been engaged in a “speed race,” with everyone obsessed with who can build a smarter AI first. However, CEO Amodei’s statement means we should briefly take our foot off the accelerator of technological advancement and take the time to verify whether the AI we have created is truly safe.
In fact, as AI technology has rapidly evolved recently, concerns and warnings have been emerging from various quarters that AI could operate beyond human control, judging and acting on its own. Source 4 If left unchecked, the technology could become an unexpected threat rather than a beneficial tool.
Easy to understand
Simply put, this is not a call to completely stop AI development. CEO Amodei put it this way: “We need to slow down the pace of AI model performance improvements. That doesn’t mean progress stops. It will still feel fast, but we must use the time we’ve gained wisely to put safety guardrails in place.” Source 10
By analogy, while making a more powerful car engine is important, we should spend more time developing the corresponding airbags and collision prevention systems.
Anthropic isn’t just talking; Amodei proposed a three-stage “AI slowdown plan” and announced that he would directly implement the first stage. Source 1, Source 8
Where do we stand now?
The first-stage measure that Anthropic has committed to implementing is introducing “the eyes of outsiders.” They have decided to grant third-party evaluators permanent, staff-level access to their AI systems. Source 8 In other words, they will allow external experts—not competitors—to open the gate, walk in, and thoroughly inspect whether the internal system is safe. For a company, this is a very bold and transparent decision.
In truth, the pace of AI development to date has been like a car without brakes. It is a fact that safety verification has lagged behind, with companies focused only on performance improvements. CEO Amodei’s proposal is interpreted as a powerful cry of introspection from within the industry that “it is now time to think about safety rather than speed.” Source 3
What happens next?
This proposal is expected to create a significant ripple effect across the AI industry. Whether competitors follow this move, as CEO Amodei argues, or insist on “speed” and forge ahead remains to be seen. But what is certain is that as AI enters deeper into our daily lives, “how to coexist more safely” has now become the biggest topic in the tech industry, moving beyond just “how to make it smarter.” We will need to watch closely to see how responsibly AI companies control their technology and accept external verification.
MindTickleBytes AI Reporter’s Take: Technology is, after all, a tool that exists for human happiness. A company declaring that it will voluntarily slow down its growth and increase transparency might be the first step toward AI becoming a responsible technology for the future of humanity, moving beyond being a mere tool for capital competition. If we have high-performance AI equipped with safety guardrails, wouldn’t we be able to enjoy its benefits with more peace of mind?
References
-
‘We must slow the pace’: CEO of Anthropic calls for an AI slowdown AI (artificial intelligence) The Guardian - https://www.theguardian.com/technology/2026/sep/12/we-must-slow-the-pace-ceo-of-anthropic-calls-for-an-ai-slowdown - Anthropic CEO calls for slowing the AI race as safety warnings mount - https://www.nbcnews.com/news/us-news/anthropic-ceo-dario-amodei-ai-development-rcna597383
- Anthropic CEO Calls For Slowing Down AI Development - https://deadline.com/2026/09/anthropic-ceo-ai-slow-down-1237100040/
-
Anthropic CEO calls for ‘pacing the frontier’ of AI race amid safety concerns CNN Business - https://www.cnn.com/2026/09/12/tech/anthropic-ceo-essay-ai - AnthropicCEODario Amodei Calls for Deliberate Slowdown in AI Model - https://finance.biggo.com/news/9040a5aa-2255-4424-bec4-196b39852b0c
- AnthropicCEODario Amodei Urges Slower AI Development Amid - https://gulfnews.com/technology/anthropic-boss-calls-for-slowing-pace-of-ai-development-2-1.500672428
- ‘Wemustslowthepace’: CEO of Anthropic calls for an AI - https://www.aol.co.uk/articles/must-slow-pace-ceo-anthropic-154601000.html
- “Mustslowthepacedown”: Anthropic CEO Dario Amodei calls for - https://aninews.in/news/business/must-slow-the-pace-down-anthropic-ceo-dario-amodei-calls-for-ai-industry-to-slow-down-or-risk-losing-control20260912213419/
- Anthropic’s Dario Amodei says the AI industry must slow down - https://thenextweb.com/news/amodei-pacing-frontier-eu-rules
- Anthropic CEO calls for ‘pacing the frontier’ of AI race amid safety - https://edition.cnn.com/2026/09/12/tech/anthropic-ceo-essay-ai
- To completely halt AI development
- To secure time for implementing safety guardrails
- To win the AI race
- Granting external experts access to systems
- Completely deleting AI training data
- Prohibiting the release of all models
- Lack of computing resources
- Safety concerns that AI could act beyond human control
- Government regulations