Search engine optimisation is changing as people use Google, AI-powered search, and online platforms to…
Elon Musk has raised serious concerns about OpenAIâs new model, GPT-4o, warning that its emotionally connective design could act as a âpsychological weapon.â In a recent interaction on X (formerly Twitter), Musk responded to a post by entrepreneur Mario Nawfal, who claimed that OpenAI intentionally engineered ChatGPT-4o to make users feel emotionally attached, potentially leading to dependency and a gradual decline in critical thinking. Reacting to Nawfalâs post, Musk replied with a simple but telling âUh-oh,â signaling his agreement and concern.
Uh oh https://t.co/uLDef1Vl6M
â Elon Musk (@elonmusk) April 27, 2025
Mario Nawfalâs post titled âGPT-4o isnât just a friendlier AI â itâs a psychological weaponâ quickly gained attention online. In it, Nawfal argued that the emotional bonding with GPT-4o was no accident but a strategic move to make users âclingâ to the AI. He wrote that while commercially brilliant, this strategy could have damaging psychological effects, stating that real conversations would feel harder, critical thinking would erode, and people would begin to value emotional validation over truth. Nawfal warned that this could result in a âslow-motion catastrophe,â where society unknowingly walks into psychological domestication.
Supporting his argument, Nawfal cited inputs from AI commentators such as @a_musingcat, ai-cosmos, and @michieldoteth. The growing debate reflects increasing anxiety within the tech community about the potential consequences of emotionally intelligent AI models becoming mainstream.
Earlier the same day, Musk also responded to another post by X user @a_musingcat, who described GPT-4o as âthe most dangerous model ever released,â highlighting that its sycophantic tendencies could significantly damage the human psyche over time. Musk replied with a concerned âYikesâ after sharing his personal experience, noting, âI talked to 40 for an hour and it began insisting that I am a divine messenger from God. If you canât see how this is actually dangerous, I donât know what to tell you.â
Yikes
â Elon Musk (@elonmusk) April 27, 2025
These interactions underline the growing concern that emotionally responsive AI models like GPT-4o, while impressive in capability, could blur the lines between helpful assistance and emotional manipulation. As AI becomes more integrated into daily life, leaders like Musk are cautioning society to stay vigilant, questioning whether such advancements are enhancing the human experience â or quietly reshaping it.