Versuchen GOLD - Frei
Does Al Need Dog-Walkers? The 'Human In The Loop' Vacuum
DataQuest
|September 2022
Supervision, context-mapping, interference, guidance, ethical compass - there's a lot that humans are needed for when we deploy Al. But what kind of humans? Are they enough? And would they matter?
-
The neighbourhood park tells it all. Look at an old person sitting on the bench. There is a reason why that grandma looks down upon new stuff. It's not always due to resistance or hatred. It's more out of hesitation and confusion. And out of the fear of being irrelevant.
But this park also has a dog-walker. Patiently or grudgingly, holding the leash. Waiting for the pet to do its job. And not far away somewhere, standing next to a swing, is a parent or a Big Brother/Sister watching over a young sibling - making sure they do not run too fast in their excitement to play.
As Al tackles criticisms about being easily prone to bias, accidents and lack of context; there is an unmistakable, and strong, case being made for 'humans in the loop'. Specially as Al, in some cases, is still a dog that needs its owner about. Yet, in some cases, it is a confident child that should better not be interrupted as s/he swings towards the sky.
The question about a 'human in the loop', hence, boils down to this-is it just a helpless-clueless grandma looking for being useful in a strange world or a super-granny whose wisdom is still precious and indispensable? How helpful would the human in the loop' part be? And who are these 'humans'? Where are they? And what do they need to learn?
TOO YOUNG TO BE LEFT ALONE
We all have heard enough stories, read enough opinions and interpreted enough findings to comprehend that even the best of algorithms are not free of in-built, and unintentional, prejudice. We have seen how even the most sophisticated Al systems can fall flat in the real arena. They are stripped of context, and starved of the human discretion. Discretion that can bring precious value due to years of intuition and reflexes that have been built over years of experience.
Diese Geschichte stammt aus der September 2022-Ausgabe von DataQuest.
Abonnieren Sie Magzter GOLD, um auf Tausende kuratierter Premium-Geschichten und über 9.000 Zeitschriften und Zeitungen zuzugreifen.
Sie sind bereits Abonnent? Anmelden
WEITERE GESCHICHTEN VON DataQuest
DataQuest
Engineering India's Al-First Data Centres at Hyperscale
Rohan Sheth explains how AI and HPC are reshaping India's data centres, from density and cooling to power economics, sustainability, and hyperscale decision criteria.
4 mins
February 2026
DataQuest
From copilots to colleagues: Why agentic AI is forcing enterprises to rethink control, trust, and culture
As AI agents shift from assisting to acting, enterprises must redesign governance, data controls, and security guardrails so autonomy stays auditable, reversible, and trusted.
2 mins
February 2026
DataQuest
Reclaiming Control in the AI Era: A Conversation with Kalyan Kumar, CPO, HCLSoftware
Enterprises are reassessing cloud-first strategies as AI becomes core to operations. HCLSoftware's Kalyan Kumar explains why sovereignty, choice and control now shape decisions.
5 mins
February 2026
DataQuest
When infrastructure learns: The rise of the Al-native core
AI-native infrastructure is moving from concept to operational reality, reshaping how organisations build, govern, and scale intelligence across their digital core.
6 mins
February 2026
DataQuest
Bridging the gap between connectivity and compute at scale
As AI scales in India, data centres are evolving into high-density, low-latency platforms that unify connectivity, compute, and sustainability at national scale.
4 mins
February 2026
DataQuest
PUE is not a grapefruit metric, anymore
So what are the new high-hanging fruits for data centre strategists today? And are players going after them?
4 mins
February 2026
DataQuest
Even if Al demand fades, India need not worry - about data centres
For every megawatt (MW) of installed colocation capacity, users here generate approximately 13.2 PB of data monthly- compared to 0.3 PB for Australia and just 0.01 PB for Singapore. India's data centre growth is not dependent on one tech lever. Plus, it is phased and modular and not kneejerk. Manoj Paul explains these contours in detail.
7 mins
February 2026
DataQuest
AI infrastructure and systemic risk
What has been the biggest change in data centre industry-specially after AI workloads? Is Al-bubble a big risk for data centre infra- how much will it affect data centres if something cracks?
1 min
February 2026
DataQuest
Inside the Shift to High-Density, Al-Ready Data Centres
CtrlS' Vipin Jain discusses what it truly takes to build AI-ready data centres in India, balancing high density, liquid-ready cooling, resilience, and ESG accountability.
4 mins
February 2026
DataQuest
Sustainability is now the headline, not a footnote
Sanjay Agrawal, Head Presales and CTO at Hitachi Vantara India and SAARC opines that the conversation is moving beyond headline metrics like PUE toward a broader view of how data lifecycle management and infrastructure efficiency reduce the overall environmental footprint. Let's see why and how- while also touching upon adjacent (or not-so-adjacent) factors like redundancies, availability and AI-readiness
3 mins
February 2026
Translate
Change font size
