Rhik Samadder, a self-declared AI sceptic, has begun a six-week experiment for the AI for the People series by turning to ChatGPT as a therapist. In his first session, he typed a detailed account of his struggles caring for his 82-year-old mother, describing exhaustion, resentment and guilt as an only child. “I snap, and shout, then struggle with guilt,” he wrote. “I’m resentful, irritable, and I love her so much. Please help me.”
The chatbot responded with a seven-point care plan, including a triage system for tasks and tips to lower the emotional temperature of interactions. It also offered reassurance, telling him: “You’re not failing. You’re carrying a load that would flatten most people.” Samadder admitted the response moved him to tears, but felt ambivalent about receiving compassion from a machine, likening the sensation to MDMA.
He had previously stopped human therapy after falling in love with his therapist, and found the AI’s approach comparable to cognitive behavioural therapy – helpful but incomplete. The following day, he tested a Jesus AI chatbot trained on religious texts, asking about open relationships and whether to have children. It quoted Hebrews 13:4 and advised seeking God’s guidance, but lacked the wit of his former therapist.
Despite his reservations about pattern-predicting software with no accountability, Samadder acknowledged that the experience was “wonderful” and calming. He noted, however, that certain forms of loneliness and unbearable news should be held in human time and relationship, not processed in seconds on a screen.
