{"id":1610,"date":"2026-10-25T03:55:12","date_gmt":"2026-10-25T03:55:12","guid":{"rendered":"https:\/\/aipathfinder.org\/?p=1610"},"modified":"2026-10-25T03:55:13","modified_gmt":"2026-10-25T03:55:13","slug":"an-open-letter-to-dario-amodei-and-sam-altman","status":"publish","type":"post","link":"https:\/\/aipathfinder.org\/index.php\/2026\/10\/25\/an-open-letter-to-dario-amodei-and-sam-altman\/","title":{"rendered":"An open letter to Dario Amodei and Sam Altman"},"content":{"rendered":"\n

We have all been brought to a precarious edge of the cliff position by big tech companies due to their failure to take a practical, balanced and humane view of the risks posed by AI to human beings. The big tech companies are looking in the wrong direction because they are involved in an ego, greed and power struggle out of which they are unable to extricate themselves.<\/p>\n\n\n\n

The greed and hunger for power and domination of the big tech company founders and their venture capitalists have led to the choice of AI risk management in a way that has dismally failed to consider the effects of their actions on the rest of humanity.<\/p>\n\n\n\n

The rest of humanity constitute 8.4 billion people spread across the globe, in far less advanced and poor countries who have equal rights of survival in this world. They have an equal right to peaceful lives as a founder or a venture capitalist in Silicon Valley and by extension in decisions which disturb the peaceful existence and survival of humanity. The blinkered view of a few people should not be the criterion deciding the fate, peaceful lives and existence or not of 8.4 billion people of this planet.<\/p>\n\n\n\n

The decision in question about the direction, progress and risk management of frontier AI models is the single most cataclysmic decision than any other at any time. This exceeds the threat to the human race posed by competing threats like wars, Chemical, Biological, Radiological, and Nuclear (CBRN), pandemic and climate change.<\/p>\n\n\n\n

The right ways of measuring, monitoring and controlling human centric AI risk are not even a topic of discussion today.<\/p>\n\n\n\n

Being a technology, the definition of risk of AI systems should be viewed from an engineering perspective which provides explainability, promotes understanding and acceptability among people. The developers must be responsible and accountable for their work. These factors have been completely missing since the discovery of neural networks and the transformer architecture. Black box syndromes have been wished away as if they were non-issues! Human requirements and societal needs which drove technology development for decades have been increasingly ignored since the development of social media in the early 2000s and more so into the AI frontier model era in 2023. Software engineering discipline which ensured engineering quality and safety and the meeting of specific defined objectives which could be transcribed into a contract with a customer is missing altogether. Why is there no software engineering principles applied to AI design and development today? There is no justifiable answer to this question except that the big tech companies have decided to look the other way. And what about responsibility and accountability? Do they not feel responsible and accountable for their work and the damage they cause? Or is it because governments are not sure what to do and are sitting on the fence?<\/p>\n\n\n\n

The correct way to measure AI risk is to study the different ways in which AI will impact human lives and then decide whether we want it to embrace it or not for that particular use. My<\/p>\n\n\n\n

paper<\/a> explains this approach. Asking questions like whether we want AI to automate recruitment decisions or eliminate our jobs. The risk indicator level of the measure should be the deciding factor in whether the use case deserves to be automated with AI or not. This is the sane path.<\/p>\n\n\n\n

Innovation may continue but must not be operationalised. Innovation and research may continue for obtaining scientific, engineering clarity and solutions, while ensuring explainability, responsibility and accountability in production systems. Till such time as the above factors are achieved and measurement of human centric risk falls below specified levels for a use-case, newer and more advanced models must not go into production environment. Software engineering principles must be mandatory for AI development. The process should start with requirement and needs of customers which will drive AI use case innovation and development. Not the other way round where some inexplicable advance is spotted and pursued blindly as is happening today and is the perfect recipe for disaster for all of us. This change in direction does not take away anything from your work while at the same time giving it sanity and legitimacy which is sorely lacking in the AI industry today.<\/p>\n\n\n\n

The above approach is in stark contrast with the proposed Anthropic proposal to the US Government setting out an imminent risk of $1 bn in damages or killing or serious harm of 50 people for governance to kick in! Every single human life matters and setting a bar like Anthropic has done is like playing God the Creator because it can dispense with human lives and large financial losses with no responsibility or accountability whatsoever.<\/p>\n\n\n\n

My questions to Dario Amodei are these \u201cSir, can you create life? If not, who or what gives you the right to decide to take 50 lives to support your projects?\u201d \u201cWhy don\u2019t you donate $1 bn from your own pocket for the damages caused instead of writing it off?” Sir, it is only in that manner can we become responsible and accountable.<\/p>\n\n\n\n

If we drop the foundations of our scientific and engineering knowledge and technology, we are sure to have the house crumbling down sooner than later. Except in this case, it is not the house but the whole of humanity and our world.<\/p>\n\n\n\n

AI is very different from any other technology when it comes to the power and scale of denial and destruction. It is the only technology which has the power to impact each and every person on the planet destructively. Even nuclear bombs and chemical, radiological and biological warfare well find it hard to destroy entire populations.<\/p>\n\n\n\n

The current alignment debate is absurd. Open AI does not know what 700 of its agents did and even after 3 months it is still trying to understand it. This clearly shows AI is faster than us and we will always lag behind provided<\/strong> we get to know what is happening, which we do not<\/strong>. This also means that we cannot track what is happening today with agents that break free. How can we then hope to understand or track, access, monitor, and control what happens when AI agents become more intelligent, faster and more powerful? In this precarious situation why are we focusing on alignment? There are straightforward reasons already to stop further development and new release of frontier models. Open AI itself has professed its difficulty when it reported yesterday that its monitoring instructions were removed by an agent being tested.<\/p>\n\n\n\n

There are many people in the world who are not foolish enough to keep banging their heads against the wall till it breaks. But clearly you both are not among them. You also forget that you are banging all the heads of every human being on the planet against the wall by your selfish actions.<\/p>\n\n\n\n

The bottom line here is that there will be support for AI innovation provided its effects can be controlled in deterministic rather than probabilistic ways and the safety, security and peaceful lives of humans is not affected. This points to the need for engineering basis to underpin AI development which is missing today. Risk must be measured as the AI impact on each human being for every use case. Every life matters. We should be led by our intelligence and not forced into corners by the greed and power-driven intoxication as big tech has already done.<\/p>\n\n\n\n

You both do know very well that there is very limited time to extricate ourselves from this dangerous situation if you do not heed this letter. You both also have the connections and power to set this right. I sincerely hope you will use your strengths to do what is right for all of us also. All it requires is the courage to decide and stand firm. No power on earth can shake the decision made by the mind of a human being. Thank you both.<\/p>\n\n\n\n

About the Author<\/a><\/p>\n\n\n\n

Disclaimer:<\/em> The opinions expressed in this article are personal opinions and futuristic thoughts of the author. No comment or opinion expressed in this article is made with any intent to discredit, malign, cause damage, loss to or criticize or in any other way disadvantage any person, people, company, government, country or global and regional agencies.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

We have all been brought to a precarious edge of the cliff position by big […]<\/p>\n","protected":false},"author":4,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[27,17,18,42,26,1],"tags":[30,32,33,34],"class_list":["post-1610","post","type-post","status-publish","format-standard","hentry","category-ai-analysis-and-design","category-future-of-ai","category-loss-to-humans","category-human-rights","category-opinion-piece","category-reg","tag-ai-analysis-and-design","tag-human-centric-ai-2","tag-responsible-ai","tag-trustworthy-ai"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/posts\/1610","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/comments?post=1610"}],"version-history":[{"count":3,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/posts\/1610\/revisions"}],"predecessor-version":[{"id":1615,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/posts\/1610\/revisions\/1615"}],"wp:attachment":[{"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/media?parent=1610"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/categories?post=1610"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aipathfinder.org\/index.php\/wp-json\/wp\/v2\/tags?post=1610"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}