Which is better -- real or artificial intelligence?

The question of which is better, artificial or real intelligence contains several concepts that need unpacking first in order to make a reasonable answer.

First what exactly do we mean by intelligence?

There are a number of qualitative characteristics that seem to constitute intelligence. The ability to make sense of various inputs from the environment as well as remembering and learning from past experience are considered signs of intelligence. Also, the ability to make sense of ambiguity or seemingly contradictory inputs and dealing with perplexity through rational inference are components of intelligence. Intelligence means using a reasoned or rational thought process to respond and react successfully to new situations or scenarios. Intelligence is complex and adaptive. Intelligence is recognizing and making judgments about the relative importance of different elements within a situation to arrive at reasonable and effective responses.

Second, what do we mean by artificial intelligence? (I am assuming that 'real' intelligence is that exhibited by human agents).

Artificial intelligence, also called machine intelligence, is behavior performed by a computer system that if done by a human would be considered intelligent behavior. AI differs from the typical computer system in that it focuses on symbolic as opposed to numeric manipulation and utilizes heuristic processing as opposed to algorithmic processing. Heuristics are basically a form of intuitive knowledge, often called rules of thumb. Rather than a specific and rigid algorithm, heuristics are flexible and adaptive. AI's represent knowledge symbolically and manipulate that knowledge using heuristics thereby becoming capable of responses that appear in context to replicate real intelligence. Perhaps the best example is the chess program Deep Blue developed by IBM which beat Chess Grand Master Gary Kasparov. (Chess has long been considered a yardstick of measuring intelligent behavior and seems to be a perfect replicator of symbolic heuristic manipulation.)

Lastly, what do we mean by better? Better in what sense?

If it is simply a matter of preference, not surprisingly we humans would appear to prefer real (or human) intelligence. But there are a number of areas in which AI is arguably superior. Because artificial intelligence operates within a computer platform, it is well documented in a way that real intelligence is not. It is also internally consistent and thorough. It does not suffer from fatigue or have bad days. AI tends to execute tasks much faster than a human could and perform those tasks with a higher degree of accuracy. It also has a characteristic of permanence that real intelligence unfortunately lacks, as well as being easier to duplicate and disseminate. For certain tasks, AI is considerably cheaper to develop and deploy. But even with all of these advantages, there are still areas in which it is inferior to real intelligence. AIs are as of yet not naturally creative and lack inspiration. There are no AI Mozarts or Picassos yet. They are also inferior to human intelligence with regard to sensing the environment directly and adapting quickly. AIs are inferior to human jet fighter pilots.

So it would probably be correct to say, each is good at certain things and not so good at other things.

Reference

Turban, E., Sharda, R., & Delen, D. (2010). Decision Support and Business Intelligent Systems (9th ed.). Upper Saddle River, NJ: Prentice Hall.