#Business #Approaches
Escape from agentic loop · “Human-in-the-Loop is the new doom scrolling.” https://ilo.im/16ctqk

_____
#Orchestrating #AI #Agents #OpenClaw #HITL #Work #Workflows #Design #Development

Santiago sold his company after 16 years and built an agent to handle his job search: Career-Ops, open source on GitHub.

740+ offers evaluated. 354 tailored CVs generated. 68 applications sent. Outcome: Head of Applied AI.

Crucially — it's HITL by design. The system automates analysis. The human makes every submission decision. Not spray-and-pray.

Built on Claude Code + Playwright. 12K stars in two days.

github.com/santifer/career-ops

#OpenSource #AIAgents #JobSearch #HITL

Claiming that AI usage is safe and responsible if there is a human in the loop is a deception. AI may be “just a tool” but the person responsible for using it is not the “human in the loop”, but the person responsible for the loop. #AI #NoToAI #HITL

http://ericlawton.org/2026/03/20/2645/

There are people who claim that LLMs and other forms of AI™[1]are merely another tool which humans can use for good or bad.

There are others who say that it is OK to use AI™ for various purposes, as long as there is a “Human in the Loop” (HITL). These contradict each other.

Human-in-the-loop refers to a system or process in which a human actively participates in the operation, supervision or decision-making of an automated system. In the context of AI, HITL means that humans are involved at some point in the AI workflow to ensure accuracy, safety, accountability or ethical decision-making.

IBM

The contradiction shows best in examples. When you use a hammer to drive a nail, this is a straightforward use of a tool.

When you have a process in which several of the steps are machine operations and others are human, then the human is not using the machine as a tool, the overall system is using both the machines and the humans as tools.

For example, if a military system uses AI to analyze some features of the battlefield to find a target, then prepare a weapon to hit that target, but does not fire the weapon until a human soldier presses the “fire” button, it is not that soldier who used the AI as a tool, but the military unit that configured the whole weapon-AI-soldier system that is using the weapon, the AI and the soldier as tools to achieve its objective.

This poses a high risk of deception, where the purpose is to avoid moral responsibility altogether or at least to shift it to the soldier who is not merely following orders but given little opportunity to make the moral choice themself and so make the [Nuremberg defense]https://en.wikipedia.org/wiki/Superior_orders moot.

If the soldier makes the decision to fire without the use of the AI altogether, based solely on their own senses they are clearly responsible. Even if there is a lot of smoke and confusion, they are responsible for incorporating that. “I couldn’t be sure but I fired anyway” is only a weak excuse for killing a child instead of an opposing soldier, with some credit to be given if there was a high risk of the soldier being killed if they hesitated and none if they were flying above with little risk to themself.

When the AI is involved, then if the soldier has access to all the data that the AI has available, possibly including radar, infrared and images from other vantage-points, and the time to assess it all, there was no point in using the AI. Otherwise, the soldier must assume the the AI has more information or faster processing speed and must make a decision biased towards that assumption. They are no longer fully accountable for that decision. They have been deceived into making the decision.

This is increasingly true as the AI makes “better” decisions—the ones the soldier would have made given the time and information. And increasingly true the more voluminous the information the AI has access too and the shorter the decision period, the less real is the claim that soldier is really in the loop at all.

Summary

AI may be “just a tool” but the person responsible for using it is not the “human in the loop”, but the person responsible for the loop.

Th #HITL Accountability gap. Because #AI is never responsible.

https://smallwarsjournal.com/2026/03/11/human-in-the-loop/

Waymos "selbstfahrende" Autos sind oft nur Typen auf den Philippinen

https://winfuture.de/news,156707.html

> Waymo muss für seine Robotaxis weiterhin auf ein menschliches Sicherheitsnetz zurückgreifen. Wie eine Anhörung im US-Senat offenbart, helfen Mitarbeiter von den Philippinen bei komplexen Verkehrslagen nach - ein Kratzer im Lack der Google-Tochter.

Alles getürkt.

#AI #slop #hitl

"Australia remains the only liberal democracy not to have adopted its own constitutional or statutory bill of rights. Instead there is a patchwork of local laws and international agreements...The Australian commitment to human rights is always subject to political whim."
>>
https://www.theguardian.com/commentisfree/2026/jan/18/for-all-the-talk-of-australian-values-our-rights-as-citizens-and-humans-remain-fragile

Hartmut Rosa, Situation and Constellation, On the Disappearance of Human Agency
"Particularly in our professional lives – but increasingly also in our leisure time – our decision-making is prescribed in advance by guidelines and forms, algorithms and apps, right down to the smallest of details. The role of situational, sensitive consideration and judgment is being replaced by the constellation-based machine logic of execution, which we wield on a daily basis. »Agree«/»Reject« – this is how agents of action turn into executors of actions."

"...When spheres of discretion and judgment disappear and the creativity of human agency is eliminated from everyday practices of execution, we feel an increased sense of powerlessness. And as our power of judgment dwindles, so does our energy to act."
>>
https://www.suhrkamp.de/rights/book/hartmut-rosa-situation-and-constellation-fr-9783518588338

#HumanRights #citizens #dehumanisation #BillOfRights #HumanAgency #HumanAction #PowerOfJudgment #machine #algorithm #robodebt #HumanInTheLoop #HITL #HeterogeneousEnsemble #assemblage #apparatus #ToDoList #governance

인간이 AI 감독하는 시대 끝났다: Agent-in-the-Loop의 등장

https://aisparkup.com/posts/7451

💼 Alignerr @alignerrai

Labelbox’s Alignerr: a contractor marketplace for RLHF, eval, and data-label tasks — rates shown per project.

https://www.everydev.ai/tools/alignerr

#RLHF #DataLabeling
#HITL #HumanInTheLoop #ModelEvals
#AIGigs
#DataTasks