Last week, the world seemed to inch closer to midnight. 

Reports emerged that China had tested a nuclear-capable hypersonic missile. Yesterday Mark Milley, the US's top general, confirmed them.

Hypersonic missiles can travel at five times the speed of sound and evade air defence. The US and Russia have conducted several tests.

The missiles could mark the launch of a new Cold War. I don't know if it's quite a Sputnik momentA period of Cold War fear in Western nations following the Soviet Union's 1957 launch of Sputnik 1, the first man-made satellite. ," said Milley, "but I think it's very close to that."

According to a book by three of the world's greatest authorities, this is the least of our worries.

When physicist Robert Oppenheimer witnessed the first atomic bomb test in 1945, he quoted the Bhagavad GitaOne of the holy scriptures of Hinduism, which contains many of the religion's main doctrines. : "Now I am become Death, destroyer of worlds." Humanity had entered a new age haunted by the shadow of mutual assured destructionA military strategy in which an attack by one power would be met by a response from the other that would lead to the annihilation of both..

The Age of AI, by diplomat Henry Kissinger, computing professor Daniel Huttenlocher and former Google head Eric Schmidt, places humanity on the brink of a more unpredictable moment.

The trio argues that artificial intelligence (AI) is the most dangerous technology ever developed. While nuclear weapons are expensive, AI can be developed easily.

AIs have already exhibited remarkable achievements. They can translate languages, recognise faces and detect cancer. 

Armies have begun to use automated weapons over soldiers. In 2018, a drone almost assassinated Venezuelan President Nicolas Maduro. It costs about $1,000 to build such a device. Tech billionaire Elon Musk Tweeted in 2014 that AI is "Potentially more dangerous than nukes."

In some fields, AI has surpassed humans. Google's chess-playing AlphaZero has defeated programmes based on grandmastersA title awarded to the world's greatest chess players, second only to the World Champion. 1948 players have been granted the honour.. Another AI algorithm defeated a US Air Force pilot in a dog fight.

This presents a terrifying question. Do things learnt by AIs count as human knowledge? If not, we might no longer be the most intelligent creatures. The rise of AI could see our share of knowledge shrink, as machines come to know things we cannot comprehend.

Others point out that AI remains limited. For a robot to learn the difference between an apple and an orange, it needs to be exposed to a vast amount of data. A child can do the same very quickly, then apply that learning to other situations.

There are different kinds of intelligence. AIs might never have the emotional intelligence acquired to emphasise, communicate and work together. 

They also lack creativity. An AI can be programmed to make artwork. But they will never be able to capture the artistic intelligence we see in the paintings of RembrandtAmong the Dutch master's works are a series of penetrating self-portraits. or hear in the music of BeethovenGerman composer (1770 - 1827) whose intense, dramatic work revolutionised classical music..

Yes. And technology is to blame. The internet has granted us access to knowledge with a flick of a finger. Our attention spans have shortened. This will only get worse as we come to rely on AI.

No. AI might be able to calculate better than any human. But it is humans who create them, set their purpose and reap the rewards. From the abacus to the calculator, people have always used tools as shortcuts. Today's AI is no different.