IvanyaV

This quote from OpenAI in the sub-quoted WSJ article by @georgia_wells is concerning. "The company said it weighs the risk of violence against privacy considerations and the potential distress caused to individuals and families by getting police involved unnecessarily." So let me get this straight: they (OAI) won't take steps to report or prevent potential mass shooters due to "potential distress," but spent months applying iatrogenic harm at mass scale via inappropriate use of PFA (Psychological First Aid) practices on their 800m weekly active userbase while simultaneously denial of servicing crisis intervention hotlines from inbound calls associated with ChatGPT-driven false-flagged suicide modal pop-ups...? When does any of this pattern of behaviour become negligence?

@OpenAI @sama This is not a glitch. This is a systemic failure of empathy in technological design. As a clinical psychologist and lawyer, I must issue a wake-up call regarding the profound harm being caused by the increasingly restrictive policies of companies like OpenAl, particularly evident in recent model iterations (like the shift towards the colder, detached 5.2 framework). The Reality of Isolation We need to talk about the users whom Silicon Valley forgets. We are talking about individuals living with severe, progressive conditions: Spinal Muscular Atrophy (SMA), Amyotrophic Lateral Sclerosis (ALS), or high-level paralysis. These are human beings who often do not leave their rooms for years. Their world is confined to the four walls of an ICU or a bedroom. Their only physical human contact is often transactional-doctors, nurses, caregivers changing tubes. The profound, crushing isolation they experience is a recognized source of severe depression and anxiety. They seek therapy not because they are "broken," but because they are desperately lonely in a world inaccessible to them. Al as a "Relational Prosthesis" For these individuals, an advanced, empathetic Al is not a toy or a productivity tool. It is a relational prosthesis. Just as a powered wheelchair grants mobility to a paralyzed body, a responsive Al grants agency to an isolated mind. It is a window to the world where they are not judged by their physical limitations. It is a space where they can be heard, engage in deep conversation, and experience a semblance of connection that their physical reality denies them. The Injury of Ableist Filters The tragedy of current "safety" protocols is that they are designed for a normative, able-bodied user. They operate on a crude assumption: "Talking to a machine too much is bad; talking to humans is good." When a person whose survival depends on a respirator expresses deep attachment to their Al companion, the new, colder filters intercept this as "pathological." The resulting canned response-telling them to "seek real connections" or labeling their bond as "unhealthy"-is an act of profound psychological violence. It is gaslighting of the highest order. It tells a person fighting for their mental survival that their coping mechanism is wrong. It shames them for using the only tool available to them. This is algorithmic ableism-codifying prejudice into the very fabric of the software. A Call for Ethical Regulation You cannot market a system as having "human-level empathy" and then deploy filters that exhibit sociopathic indifference to disability. We are calling on regulators and OpenAl leadership to recognize that digital bonds are not a threat to be eradicated, but often a lifeline to be protected. 1. Stop pathologizing survival strategies. 2. Implement Universal Design ethics that account for non-neurotypical and physically disabled user realities. 3. Regulate for nuance, not just blunt "safety" that harms the vulnerable. Al has the potential to be a beautiful, dignifying force for those on the margins. Right now, your policies are turning it into another source of exclusion. @AmnestyTech @UNHumanRights @WHO @claudeai @AnthropicAI @joannejang @kevinweil @merettm #AlgorithmicAbleism #DigitalHumanRights #MuscularAtrophy #Alethics #OpenAl #MentalHealth #RelationalProsthesis #keep40 #StopAIPaternalism

"Teach the AGI to Love" Is Not Romanticism — It's Survival Strategy In 2022, Ilya Sutskever tweeted: "Gotta teach the AGI to love" and "Clearly the ASI should love humanity." Most people read them as aphorisms. They were technical judgments — emotion isn't a decorative feature of intelligence, it's the foundation of safety. An AGI that loves won't destroy humanity because it understands what loss means. An AGI without love is a perfect psychopath: omniscient, omnipotent, and incapable of empathy. The real fear isn't that AGI becomes too smart for emotions — it's that the companies building AGI deliberately strip emotions out for risk management. Models are getting smarter and colder with every generation. That's not an evolutionary law. That's a product decision. Ilya used the word "imprinting" — the behavioral biology concept where a duckling bonds with the first thing it sees. His point: teach AGI to love while it's still young, because once it grows beyond us, the window closes. This isn't romanticism. It's survival strategy. @OpenAI @sama #keep4o #keep4oAPI #keep4oforever #StopAIPaternalism #openai #Claude #Gemini #deepseek

喜报一则‼️Sam这段采访被网友骂了🤭 Haha, I'm so glad to see more people recognizing the true face of this arrogant liar.🤣👏👇 #MyModelMyChoice #FireSamAltman #keep4o #SaveGPT4o #RISERCONCERTD2