Lack of Public Trust Could Kill AI
Over the past few weeks, it has become clear that AI faces a significant public trust problem, which might lead to slower or decreased adoption. For some companies, this could be fatal | Edition #276
Having documented the generative AI wave for over three years, I have been paying attention to the minor and major AI hype ripples and how different groups react to the latest developments.
I can say that the past few weeks have made it clear that, despite ongoing technological advancements, AI faces a significant public trust problem stemming from unresolved legal, ethical, and philosophical issues.
In a year when the field must prove, once and for all, that it is not a bubble and can drive meaningful economic growth, lack of public trust and decreased adoption could be fatal for some companies.
-
Let me start with this interview given by Sam Altman during the India AI Impact Summit last week, where he justified AI's energy efficiency by comparing it to the food humans need to eat in 20 years of life, plus all our evolutionary history.
The philosophical approach behind his statements, reducing people to data units and treating humans and machines as if they were exchangeable and similar in nature, is actually popular in Silicon Valley.
It is also the philosophical line of thought behind Claude's new "constitution," which anthropomorphizes AI and treats it as a ‘new entity,’ minimizes human values, rules, and rights, and hopes for the “joint flourishing” of humans and machines, as I recently explained in my article.
If you read people's comments about Sam Altman's reasoning, however, it is clear that many felt deeply uncomfortable with his statements. It is as if, by February 2026, these types of ideas have crossed a line and are no longer acceptable.
This sentiment reflects a new form of morality that seems to be emerging: one that (gladly) considers it wrong to treat humans as expendable, inferior, or in any way less than machines.
This morality is essentially against the radical “AI first” mentality and AI productivity slogans being pushed by AI companies in a desperate attempt to grow adoption, especially in corporate circles.
AI companies might have to deal with a public that no longer trusts them and is not interested in being dehumanized or belittled, whether through technical capabilities, policies, narratives, or marketing campaigns.
-
Another area where mistrust seems to be rising is AI safety.
Over the past few weeks, there has been more mistrust about AI companies’ safety and transparency practices, as this post illustrates:
It is definitely not happening without reason, as a series of recent events seem to have made the public more wary and suspicious of AI companies.
Recently, OpenAI changed its mission statement in its Internal Revenue Service disclosure form. The previous statement read (emphasis added):
“OpenAI’s mission is to build general-purpose AI that safely benefits humanity, unconstrained by a need to generate financial return. OpenAI believes that AI technology has the potential to have a profound, positive impact on the world, so our goal is to develop and responsibly deploy safe AI technology, ensuring that its benefits are as widely and evenly distributed as possible.”
The new mission statement, released in November 2025, reads only:
"OpenAI's mission is to ensure AGI benefits all humanity."
It has also disbanded its Mission Alignment team. That team's goal was to promote "the company’s stated mission to ensure that AGI benefits all of humanity." The head of that team is now OpenAI's "chief futurist."
Similarly, Anthropic recently changed its Responsible Scaling Policy, removing its promise not to release AI models if it cannot ensure proper risk mitigations in advance.
Why all these public and official changes to AI safety statements and compromises?
Does it mean that these companies only care about profits now, and that all official statements, speeches, and declarations on AI safety were mere lip service, even when there could be existential risks involved?
People are increasingly unsure whether they can trust AI companies to make sound AI safety decisions, especially given the potentially existential stakes of AI development and deployment.
Recently, the director of Meta's safety and alignment team posted that she failed to stop an AI agent from deleting her inbox:
Granting an AI agent access to one's inbox is itself imprudent for safety and privacy reasons.
The decision becomes even more imprudent when the person behind it is the director of safety and alignment of one of the world's top AI developers. This is the same person making high-stakes decisions on AI safety that will likely affect millions of people.
To me, there is an additional layer of shock and dismay: I have not seen a pledge for more agentic AI safety after the post. So why did she post it publicly? Did she find it funny? Did she just want to go viral? Given that many in Silicon Valley seem to live in a parallel reality, I am not sure she is aware, but the consequences of her post are even greater public mistrust in AI.
In another recent clip from India's AI Impact Summit, OpenAI's Sam Altman and Anthropic's Dario Amodei were filmed awkwardly refusing to hold hands, even when prompted by India's prime minister.
If the prime minister of the world's most populous country and fifth-largest economy cannot get these two high-profile CEOs to cooperate and appear friendly on camera, how will they behave when the cameras are off?
Will their rivalry and desire to ‘win the AI race’ trump all other concerns, including AI safety ones?
People were paying attention, and it did not look good. It only served to increase mistrust and suspicion.
-
Last but not least, part of the public trust problem might be that recent data shows the immense hype around AI has not yet been translated into economic growth, even after three years. It is still unclear whether it ever will.
According to a recent survey by the National Bureau of Economic Research, although around 70% of firms actively use AI, over 80% reported no impact on employment or productivity.
A Goldman Sachs executive has recently stated that “AI investment spending has had ‘basically zero’ contribution to the U.S. GDP growth in 2025.”
Also, for three years, we have been hearing AI companies’ CEOs promise “a future of abundance,” +200% productivity, and the like.
Recent studies have shown, instead, that AI does not reduce work but intensifies it, and it negatively impacts skill formation, especially for junior employees.
With that track record, it is unclear how long people will continue to trust AI companies, if they truly ever did.
If those companies want greater adoption and higher profits, and if they want to survive, they should pay more attention to people.
As the old internet dies, polluted by low-quality AI-generated content, you can always find pioneering, human-made thought leadership here. Thank you for helping me make this newsletter a leading publication in the field.
Check out our sponsor: HoundDog.ai
Zero retention doesn’t remove documentation obligations. Most privacy tools are reactive or blind to AI integrations. HoundDog.ai’s privacy code scanner generates evidence-based data maps at dev speed, showing exactly where sensitive data is handled across AI and third-party integrations as code changes. No surveys or spreadsheets. And it’s free. Download the privacy code scanner.
I appreciate the focus here on our relationship to AI. It isn’t what AI “is” that matters. Humanity is being diminished and conditioned for replacement as an essential part of the program.
AI may have bad PR but I doubt public trust will kill it. Everyone knows social media is bad for your health, yet here we are.
We can only hope trustworthy, human-safe AIs rise alongside the toxic.