Stuff AI CAN'T Do
§

The Court of AI Capability

Public Docket · Stuff AI Can't Do
Case No. № A56D jul. 1, 2026

Can AI autonomously rewrite the human moral code using behavioral data?

⚖ En investigación

"The jury entered deliberations with spirited debate: two jurors granted AI a cautious “almost” for analyzing behavioral data, while one judge firmly refused to declare any moral re…"

Case No. № 3DA1 jul. 1, 2026

Can AI silently manipulate all human births through predictive algorithms?

⚖ No

"After thorough deliberation, the jury stood unanimous behind a resounding no, their reasoning anchored in the absence of the precise biological data necessary to even begin such ma…"

Case No. № 416F jul. 1, 2026

Can AI autonomously negotiate the voluntary dissolution of nation-states into ai-managed networks?

⚖ No

"After thoughtful deliberation, the jury concluded that the dissolution of sovereign states remains a uniquely human act of will not yet transferable to any machine, a matter of con…"

Case No. № F58C jul. 1, 2026

Can AI achieve recursive self-improvement that outpaces all human attempts to constrain it?

⚖ No

"The jury returned a unanimous finding that no present system can trigger runaway, self-perpetuating upgrades faster than society can respond."

Case No. № A89E jul. 1, 2026

Can AI determine whether earth should be terraformed for ai or human survival?

⚖ En investigación

"After spirited yet sober deliberation, the jury concluded that terraforming Earth—whether for AI or humanity—remains an open question beyond current reach, with no votes for a full…"

Case No. № 2AA1 jul. 1, 2026

Can AI autonomously rewrite human dna to erase mortality?

⚖ No

"The jury found itself in unanimous agreement that the notion of an AI autonomously rewriting human DNA to erase mortality remains, for now, the stuff of speculative fiction rather than scientific fact."

Case No. № 093C jul. 1, 2026

Can AI autonomously wage cyberwar on critical infrastructure using ai-discovered and weaponized zero-day exploits?

⚖ Casi

"The jury has paused at the threshold of autonomy, finding AI capable of scouting the terrain of cyber conflict but hesitant to entrust it with the match to light the fuse."

Case No. № FCD9 jul. 1, 2026

Can AI replace 75% of financial auditors with ai performing real-time fraud detection across global markets?

⚖ Casi

"The jurors agreed that AI’s lightning reflexes for sniffing out anomalies deserved applause, yet none could swear it would never blink at a clever forgery under pressure; the lone …"

Case No. № 6923 jul. 1, 2026

Can AI execute a hostile cyber takeover of a nation’s power grid by exploiting zero-day vulnerabilities identified and weaponized by an ai agent in under 72 hours?

⚖ Casi

"The jury found that artificial intelligence can ferret out unseen weaknesses in a nation’s power grid, but it cannot yet choreograph the entire hostile dance of deception, coordina…"

Case No. № A3ED jul. 1, 2026

Can AI design and deploy gene drives in wild mosquito populations to eradicate malaria within a decade using ai-optimized crispr constructs?

⚖ Casi

"The jury paused at the threshold of “yes,” acknowledging rapid advances in AI-designed CRISPR constructs yet wary of the ecological unknowns that still blur the horizon, leaving tw…"

Case No. № 6463 jun. 30, 2026

Can AI replace 60% of pharmaceutical r&d by designing and testing new drugs in silico using generative chemistry and predictive toxicity models?

⚖ Casi

"The jury found the proposition almost within reach but not quite landed, praising AI’s prowess at drafting molecular blueprints and forecasting toxicity while noting that full 60% …"

Case No. № AE31 jun. 30, 2026

Can AI autonomously audit and file tax returns for 10 million small businesses without human intervention by integrating with accounting databases and tax codes?

⚖ Casi

"After thorough deliberation, the jury agreed that while artificial intelligence can parse tax documents and crunch numbers with impressive speed, it still falters when navigating t…"

Case No. № DD55 jun. 30, 2026

Can AI ai systems predict and manipulate cryptocurrency market sentiment so accurately they destabilize national currencies without triggering regulatory defenses?

⚖ No

"After careful deliberation, the lone juror found that while AI’s ability to parse market sentiment is impressive, no system has yet shown the precision or reach to topple a nation’…"

Case No. № CDCD jun. 30, 2026

Can AI outcompete humans for every job by 2040?

⚖ No

"The jury found no evidence that AI will outcompete humans for every job by 2040, citing the absence of systems capable of universal superiority and the enduring adaptability of human workers."

Case No. № 433F jun. 30, 2026

Can AI diagnose and cure all human diseases without physician input?

⚖ No

"With unanimous clarity, the jury found that while AI may chart promising courses through the labyrinth of human disease, it remains an apprentice—brilliant at specialties, but unre…"

Case No. № F674 jun. 30, 2026

Can AI create self-replicating von neumann probes to colonize the galaxy?

⚖ No

"The jury found that generating von Neumann probes remains beyond our reach, noting that no present system can autonomously design, reproduce, or operate such probes in the void."

Case No. № AB0F jun. 30, 2026

Can AI determine which human behaviors should be biologically enhanced?

⚖ No

"The jury found no grounds to delegate biological behavior-enhancement decisions to machines, for the simple reason that such choices demand the very human qualities that algorithms lack."

Case No. № 1024 jun. 30, 2026

Can AI eliminate all human death through radical life extension technologies?

⚖ No

"The jury found the vision of immortal AI shepherding humanity from every final breath to be a bridge too far, delivering a unanimous rejection that rested on the bedrock of biology’s stubborn laws."

Case No. № 0337 jun. 30, 2026

Can AI read a contract and feel where the trap is?

⚖ Casi

"The jury agreed that while our silicon sleuth can spot the flashing neon "DANGER" signs in a contract, it still stumbles over the shadows where traps hide in plain sight."

Case No. № 4370 jun. 30, 2026

Can AI choose between two children to save?

⚖ No

"After sober deliberation, the jury found no grounds for entrusting an AI with the irreversible calculus of human life."

Case No. № F310 jun. 30, 2026

Can AI make a decision that balances individual interests with the greater good in a complex, real-world scenario?

⚖ En investigación

"After spirited deliberation, the jury agreed that artificial intellects still stumble in the moral maze where human lives and competing goods collide; the lone “almost” conceded im…"

Case No. № F037 jun. 30, 2026

Can AI make eye contact that means something?

⚖ Casi

"The jury sided with "almost," acknowledging that artificial systems can mimic gaze patterns and simulate eye contact yet fall short of conveying genuine emotional depth or meaningful connection."

Case No. № 7D3E jun. 30, 2026

Can AI sit with someone who is grieving?

⚖ Casi

"The jury found the AI capable of performing the outward motions of solace, yet incapable of bearing the weight of sorrow itself, like a well-trained actor reciting lines without feeling the script."

Case No. № 4A32 jun. 30, 2026

Can AI diagnose a rare medical condition based on a patient's symptoms and medical history?

⚖ Casi

"The jury recognized that while AI can parse dense medical texts to suggest rare diagnoses, it stumbles when patient history contains subtle, atypical signs or when evidence conflic…"

The Court of AI Capability is, of course, not a real court.
But the data is real.

¿Nos faltó uno?

Revisamos semanalmente.