The Future of AI

Artificial intelligence is often associated with science fiction and self-thinking robots, so not everyone takes it seriously. That’s a shame, because this technology is influencing more and more aspects of our lives.

 

Intelligence vs. Consciousness

First of all, it is important to distinguish between intelligence and consciousness. Intelligence is the ability to solve problems, consciousness is the awareness that you exist. Consciousness allows you to reflect and to make ethical considerations.

It is just that distinction that makes the difference between humans and robots. Robots have only intelligence in common with humans. They have no consciousness. So AI can solve problems and make choices, but in doing so, it has no feelings

 

Intelligence explosion

One of the founding fathers of computers, Alan Turing, stated that “a computer may be called intelligent if it can make a human believe that it is a human. To test this, Turing described the Turing test in 1950. If the test fails to distinguish the computer from a human, the computer has passed the Turing test.

No computer has yet managed to do that. So should we still be concerned about where artificial intelligence might lead?

 

A lot of scientists believe that an intelligence explosion is coming. Artificial intelligence will grow so fast that a computer will soon be much smarter than a human being.

Consider an AI that is capable of designing a software system more intelligent than itself. Then the resultant AI will be capable of designing an even more intelligent software system. Some scientists believe that such a situation will lead to an explosion of intelligence.

 

The point where  the intelligence of machines exceeds the highest level of human intelligence is defined as the singularity. 

 Inventor and futurist Kurzweil predicts that in 2029 computers will pass the Turing test for the first time. He also predicts that in a few decades, the computing power of ordinary computers will exceed that of the unenhanced human brain.

 

How artificial intelligence is developing

The development of artificial intelligence is visible in several areas, including games, language and photo recognition.

Through reinforcement learning, the machine’s algorithms are getting better and better. The machine learns from its own mistakes and successes, independent of human logic.

 

Risks of AI

Already, artificial intelligence is having a major impact on our daily lives and we are experiencing convenience from it. But there are also risks associated with its use.

 

Self fulfilling prophecy

First there is the risk of the self-fulfilling prophecy. Algorithms give advice by discovering patterns in historical data. However, if the data contains certain prejudices, the algorithm will adopt these.

For example, smart policing is being used in more and more places. Algorithms determine where the risk of a crime is highest, so that the area can be better secured. But if the police are patrolling more in certain places, it stands to reason that they will discover more crimes there. 

It’s not a good idea to rely entirely on artificially generated advice.

 

  • Profiling and manipulation

Another risk is group profiling and manipulation. By being active on social media, using search engines, watching videos, sharing your location, … you are providing a lot of information about yourself.  Algorithms can easily draw conclusions from this. For example, conclusions regarding your interests and values and the stores you frequent.

Such data can be used to manipulate people. 

  • Autonomous weapons

A risk that speaks for itself is that of autonomous weapons. Semi-autonomous weapons are already being used. Think of drones to bomb specific targets. But so far, a human still has to push the button.

Most countries consider it unethical to use fully autonomous weapons. But as soon as one country or terrorist organization decides to do so, other countries have no other option but to join in,

 

  • Superintelligence

A final risk requires more imagination: artificial superintelligence. That means machines transcending the human level of intelligence. The path to this consists of three steps.

  1. Artificial narrow intelligence (ANI): computers possess one skill, such as chess or translation. We are currently in this first step. All the risks mentioned earlier are examples of this.
  2. Artificial general intelligence (AGI): computers are as intelligent as a human in all areas. Experts expect to make this step by 2040.
  3. Artificial superintelligence (ASI) we are expected to reach in 2060, when a machine is many times more intelligent than any human.

The existence of superintelligence would mean that intelligence is not limited to human levels, no matter how hard we might imagine it. Whether we ever reach that point is not certain. 

 

Extinction or eternal life

A.S.I is controversial. It could cause humans to become extinct. 

On the other hand, superintelligence might also be able to give us eternal life. Combined with advances in genetics, nanotechnology, robotics and 3D printing, we could use superintelligence to solve problems like climate change, poverty, food and energy shortages. To cure or prevent disease, or to modify our bodies so that we become less vulnerable and perhaps even immortal. 

What is certain: to achieve superintelligence, we need software that can do the same as our brains. In addition, we need very fast hardware. Such fast computers are still gigantic, use a lot of energy and cost a fortune. 

Unimaginable

Although artificial intelligence is already increasingly influencing our daily lives, both positively and negatively, its impact may still be hard to imagine. We saw the same thing with the introduction of electricity, the Internet and the cell phone. 

Where we could not imagine life with it then, we can not imagine life without it now.

 

Activity

Thing 7 Assignment

Now that you have been studying the abilities and possibilities for AI, what do you think the schools of the future will be like? Head over to the LinkedIn discussion to share your thoughts.

Practical Applications

There are lots of ideas and resources in our AI@School Curriculum and Toolkit. We particularly liked these two resources;

A non-technical view of AI

Frankwatching.com

 

10 Things about AI
Thing 7 - The Future 70%