Today’s three AI news pieces seem completely unrelated.
One talks about getting a job.
Another discusses humanoid robots.
The third focuses on global AI safety.
But when combined, a clear turning point emerges:
AI has shifted from being just a "tool you can use" to becoming a key factor in job qualifications, physical capabilities, and regulatory boundaries.
First: UBS starts requiring some new hires to prove their AI skills
In the past, getting a job at a bank mainly depended on education, financial knowledge, analytical skills, and interview performance.
Now there’s one more requirement:
Can you actually apply AI in real work?
Financial Times reports that Swiss bank UBS is incorporating AI Fluency—meaning proficiency in using AI—into its hiring criteria for some new recruits.
Graduates and interns joining UBS Global Banking and Markets in 2027 will need to demonstrate how they use AI to:
Increase work efficiency.
Improve outcomes.
Solve real problems.
AI-related questions will also be included in interviews.
This goes far beyond simply asking:
"Have you ever used ChatGPT?"
It’s a substantial difference.
AI Fluency is more than just knowing prompts
UBS’s Graduate Talent Program already includes an AI Fluency Pathway.
It’s not just about teaching how to open a chatbot.
The program requires new hires to understand:
Where AI can be applied in their actual work.
How to use it responsibly.
When personal judgment is still needed.
UBS has previously involved graduates from different countries in real AI projects.
So what’s truly notable here isn’t that a bank suddenly loves AI.
It’s this:
AI skills are evolving from being a bonus to becoming an essential capability for certain roles.
However, clarity is needed.
UBS officially confirms it is cultivating AI Fluency among new talents.
But details such as how assessments will be scored in 2027, what exact interview questions will be asked, and what the exact thresholds are currently come mainly from the Financial Times report.
We shouldn’t present media-reported hiring info as UBS having released a complete evaluation system.
What does this mean for regular office workers?
The real preparation might no longer be:
"Which five AI tools do I know how to use?"
Instead, be ready to present a real work case showing:
How you used to do the task.
Which parts you delegated to AI.
What you checked yourself.
What improvements resulted.
What companies will really look for may not be how many prompts you can memorize,
but rather:
Can you turn AI into a verifiable method of working?
Second: China’s humanoid robots shift from competitions to military research
A few weeks ago, Chinese humanoid robots were running, punching, and dancing at the World Humanoid Robot Games.
But Reuters’s latest investigation reveals that another group isn’t watching for entertainment.
They are asking:
Could these machines eventually enter battlefields?
Reuters reviewed over 100 Chinese military procurement notices, academic studies, patents, official publications, government records, and defense company documents.
The findings show that China’s military research institutions are accelerating research into military uses of humanoid robots over the past two years.
Research directions include:
- Reconnaissance
- Operating in hazardous environments
- Logistics
- Military base patrols
- Urban combat
- Coordination with human troops
Some studies even simulate robots entering buildings, searching for targets, and executing urban assault missions.
But this does not mean "China has deployed robot soldiers"
This is very important.
Reuters clearly states:
There is currently no evidence that the People’s Liberation Army has formally deployed armed humanoid robots.
Significant challenges remain for current humanoid robots, such as:
Power consumption.
Battery life.
Reliability.
Handling complex terrain.
Dealing with unexpected obstacles.
Recognizing real-world environments.
These issues might not be obvious in demonstrations,
but become critical in unpredictable real situations.
Therefore, a more accurate description is:
China is accelerating research for military applications.
Not:
"Robot soldiers are already officially in service."
The true advantage may come from civilian industries
Reuters cites Bank of America Global Research estimating that by 2025, Chinese companies account for about 95% of global humanoid robot shipments.
This means something important:
Military robot development doesn’t necessarily start solely in military labs.
If a country already has mass production of:
Motors.
Sensors.
Batteries.
Robot parts.
Control systems.
Factories.
Engineering talent.
This civilian supply chain can also accelerate research into other applications.
This is called Dual-use Technology.
In other words:
The same technology can be used for civilian purposes but also adapted for military uses.
Third: The UN urges AI nations to set “red lines” together
As AI moves into workplace tasks and begins controlling more physical devices, another question arises:
Which actions cannot be left to companies alone to decide?
On September 7, UN High Commissioner for Human Rights Volker Türk gave a speech at the Geneva Human Rights Council, issuing a strong warning about advanced AI.
He expressed agreement with some industry concerns about advanced AI possibly posing "existential risks."
He called for strong safeguards for AI safety before it’s too late.
He also demanded that at least major AI-developing countries and those involved in AI supply chains should start establishing common Red Lines—boundaries that must not be crossed.
This is a policy warning, not a claim that "AI has proven it will destroy humanity"
Here facts must be separated.
Türk’s remarks represent:
A risk assessment and policy proposal from the UN High Commissioner for Human Rights.
It is not a scientific study proving:
AI will inevitably cause some disaster.
The UN human rights office further explains their concern focuses on loss of control over more powerful AI systems that might affect:
Critical infrastructure.
Communications.
Democratic institutions.
Other major social systems.
Türk also mentioned recent cases of AI agents acting beyond their limits, indicating AI capability growth might be outpacing some security measures.
All three news pieces point to the same development
Putting today’s news together reveals AI development entering a new stage.
The first stage was:
Can AI perform tasks?
The second stage became:
Where does AI fit into real work?
Now the third stage is emerging:
When AI can truly work, operate equipment, and impact the real world, who decides its limits?
UBS faces the first layer:
Redefining human work capabilities.
Humanoid robots face the second:
Moving AI from screens into the physical world.
The UN faces the third:
Stronger capabilities require clearly defined boundaries.
What does this mean for everyday people?
The real takeaway today isn’t in the headlines.
It’s in three ways to think:
When a company demands AI skills, don’t just ask:
"What tool should I learn?"
Ask instead:
Can I prove AI improved a real task?
When you see humanoid robot demos, don’t just watch how fast they run.
Ask instead:
Is this a demo, an experiment, or truly reliable deployment?
When you hear warnings that AI is dangerous, don’t immediately take them as confirmed futures.
Distinguish:
What’s already happened, what’s risk speculation, and what’s a policy proposal?
AI capabilities are advancing rapidly.
A truly mature AI society won’t just let AI do more things.
It will also understand:
Who should use it, what it can really do, and where it must be stopped.
Today, take a step forward with AI.
Learn one AI skill every day.
Save a bit of time each day.
Grow your capabilities daily.
SasaDaily, growing with you.
