Youtube Exercise » English Grammar Homework

B2-C1

Yikes.

ColdFusion

564K views 2 months ago

Warm-up Questions

When you hear the title “Yikes.”, what kind of story do you expect (funny, scary, surprising, etc.)?
What situations make you say “Yikes” in real life?
Do you think this video will be about something positive or negative? Why?

Up Next

Learn How to Talk About Yourself in English | Easy Self-Introduction for Beginners | English Podcast Mr. English Pod
Book Club Preview: The Graveyard Book – Chapter 2 Prof. Tonn
Gerund or infinitive? Ask BBC Learning English BBC Learning English
5 ways to listen better | Julian Treasure | TED TED
READING EXPLORER FOUNDATIONS UNIT 12 12B HOW WILL WE LIVE IN 2045? Royal English Centre

Start With: Word Definition

Order: Alphabetical Random

…

Yikes.

AI used in warfare + different AI for the military (00:00 – 04:49)

Transcript will load here…

1. According to the report mentioned at the start, what did the Pentagon use Anthropic’s model Claude for?

To replace human soldiers in combat
To translate military documents into Spanish
To write public press releases about defense policy
As part of an operation to capture Venezuelan President Nicholas Maduro

2. What is one key reason the military AI setup can outperform consumer AI?

Capacity is split across hundreds of millions of users
It can use dedicated hardware and concentrated compute for a single customer
It is always fully autonomous and never needs humans
It uses no data and relies only on intuition

3. What warning is given about using generative AI in life-or-death situations?

AI is always correct if it has enough compute
Humans still need to check the output because AI can get it wrong
AI should replace commanders immediately
Consumer chatbots are already perfect at targeting

Yikes.

How AI systems work in war: Maven + Claude, data, targeting, and dependence (04:49 – 10:15)

Transcript will load here…

1. What is the name of the AI system described as enabling extremely fast targeting operations?

The Sentinel Drone network
The Apollo Defense system
The Marvin Smart system
The Atlas Vision system

2. What kinds of inputs does the system gather data from (as described)?

Only public newspapers and TV broadcasts
Only soldiers’ personal phones
Satellites, surveillance, and other intelligence (with many data sources)
Only weather forecasts

3. Why did Pentagon leadership become alarmed after the Maduro raid, according to the speaker?

They decided AI was useless and should be banned immediately
They discovered the system only worked on consumer laptops
They realized they had no alternative provider and could be at risk if the software shut off or refused
They were worried the AI would increase salaries for staff

Yikes.

Fallout, OpenAI deal, public backlash, and what people can do (10:15 – 16:53)

Transcript will load here…

1. What did Sam Altman claim was included in OpenAI’s deal with the government?

A promise to make all military AI fully open-source
A guarantee that AI will never be used in war
The same safeguards Anthropic had fought for, including prohibitions on domestic mass surveillance and human responsibility for use of force
A plan to ban AI research worldwide

2. How did much of the public backlash show up, according to the video?

People bought more subscriptions to support OpenAI
A movement called Quit GPT, with users cancelling subscriptions and sharing boycott messages
A worldwide shortage of computers
A new law immediately banning AI

3. What is the #1 practical step the speaker recommends to reduce risk from AI-powered surveillance?

Move to a country with no technology
Stop using the internet forever
Buy more smart devices to confuse trackers
Get yourself off data broker sites

Yikes.

Transcript will load here…

1. What is a central idea of the video?

AI has no use in any government work
The main problem with AI is that it is too slow
Military AI is identical to consumer chatbots and can be trusted fully
AI in warfare and surveillance raises serious ethical and safety concerns

2. Why does the video emphasize “humans in the loop”?

Because AI outputs can be wrong and decisions can be life-or-death
Because laws require AI to be slower
Because humans are better at math than computers
Because dedicated hardware makes humans unnecessary

3. What is one risk of relying heavily on a single AI/software provider for military operations?

The provider will always reduce prices
The provider will prevent any surveillance automatically
The provider will guarantee perfect accuracy
The provider could refuse, shut off access, or the model could be poisoned/hallucinate

Discussion Questions

Discuss the ideas from this video

Yikes.

Should AI ever be used to “organize and prioritize” military targets? Why or why not?
What rules or safeguards do you think are necessary before governments can use AI for national security?
Is it believable that two different companies could agree on “the same safeguards” but still have very different deals? Explain.
What personal steps (if any) would you take to reduce your online data being collected or sold?