Will AI machines become humans' peers?

The pace at which artificial intelligence (AI) has been encroaching on human territory, and the resulting upheaval in society, is causing much consternation ("Charting our AI future and reflecting on what makes us human"; Jan 21).

Besides the ability to manipulate great amounts of data, AI is now capable of emulating human characteristics, such as learning, intuition, flexibility and creativity.

This was demonstrated last year by Google's AlphaGo, the AI that mastered one of the oldest and most complex games - Go - and in the process, beat grandmaster Lee Sedol.

This feat was magnified by the fact that AlphaGo's programmers do not have the capability of defeating such a high-level player themselves.

The AI has learnt to do things that its programmers cannot do and do not understand.

The basis on which AI surpasses humans in such tasks is machine learning, which these machines use to learn things without being explicitly programmed.

In other words, AI is able to improve its behaviour with experience, often by applying statistical analysis of vast stores of data.

For example, Amazon, with more than 200 million customers and three million items in its catalogue, recommends products from an analysis of a database containing millions of transactions, searches and items.

These techniques are being applied to autonomous cars, machine translators and personal assistants, among many more applications.

Will future AI also possess other human attributes such as wisdom, self-awareness and consciousness?

Philosopher David Chalmers, writing in the book Singularity, argues that it is a matter of time before AI machines are imbued with such qualities, together with their own goals, desires, preferences and values.

They will then be a peer-competitor of humankind, which can out-think us like we out-think other animals.

Perhaps, as Professor Luciano Floridi suggests, we will have to do away with the hubris with which we have placed ourselves on the pedestal of evolution, and re-evaluate our self-bestowed exceptional status.

Indeed, if we view the universe as a product of a great software, we can also be deemed the result of information.

Just as AI runs on computer codes, humans are designed by the genetic code of the nucleic acids in our cells.

Perhaps, we can even allow for the possibility of another realm in the universe - a higher reality where we are the software constructs in a simulation run by beings which have created intelligent beings (the people) who then create other intelligent beings (AI).

Maria Loh Mun Foong (Ms)

A version of this article appeared in the print edition of The Straits Times on January 26, 2017, with the headline 'Will AI machines become humans' peers?'. Print Edition | Subscribe