The future of AI

What is next for the AI? The answer will be driven by the pace, direction and funding of research, the potential good that could result and, equally importantly, the ethical issues they raise.

As with research into areas of medical science like genetic engineering and human stem cell technologies, both of which have profound potential benefits, society needs to decide how artificial intelligence should be regulated.

And those decisions should have inputs from people with diverse perspectives and expertise, including the general public, lawmakers, scientists, ethicists and medical professionals.

The discoveries scientists make can certainly have great positive impacts on our quality of life, but unintended outcomes are also possible. So it's important to consider the many possible uses and motivations for developing these technologies.

Near future

➟ Better product recommendations

➟ More AI helping with medical diagnostics and personalised medicine

➟ Brain-computer implants, bionic eyes, arms, and legs and mind-controlled exoskeletons will become more advanced, and more available 

➟ Chatbots will be commonly used by companies for customer interaction

Medium future

➟ Brain-computer interfaces will be increasingly able to enhance memory function, especially for people with traumatic or invasive memories. 

➟ Brain-to-brain interfaces will become viable between humans—sending telepathic messages not bound by location. 

➟ Automation of many jobs

➟ Personal assistants like Alexa and Siri will be commonplace and more advanced

➟ Autonomous cars on suburban roads

Distant future

➟ Cyborgs will exist 

➟ Artificial general intelligence (very distant future…)

➟ Robot servants or carers

An intelligent future

Modern computing’s high-powered technology, along with the emergence of strikingly human-like robots, would make you believe intelligent machines are not far off taking over our jobs and our homes.

But the reality is, while AI programs are sometimes better than humans at specific tasks—like pattern recognition or playing games—they’re still a long way from achieving the kind of general intelligence that we humans possess.

The mammalian brain took millions of years to develop its complex yet efficient abilities, and despite the power of modern computers, AI is not likely to match its potential for decades.

There is still much that even ‘simple’ animal brains can tell us about how information is processed and what intelligence is. For instance, many animals—like octopus, squid and even bees—display intelligence, despite not having a forebrain.

These animals can outperform AI, in terms of how quickly they can master multiple tasks, with limited time and input for learning. “Typically, deep learning networks need millions of samples before they get the hang of it,” says neuroengineer Professor Srini Srinivasan, from UQ's Queensland Brain Institute. “A bee lives for barely a month, and it’s got to learn everything within that time,” he says. “In the first week of its life it learns to forage for food, recognise species of food-bearing flowers, and navigate back home without getting lost.”

The joke is on AI: Will creativity distinguish humans from machines?

AI programs have been developed to do very human things, like write novels, compose music and create art, but can they take a joke—or even understand one? AI is more likely to be the punchline. Attempts by AI programs to write prose are often decidedly funny (to humans). 

Take this example from an AI program that was fed all seven Harry Potter books:

“Leathery sheets of rain lashed at Harry’s ghost as he walked across the grounds towards the castle. Ron was standing there and doing a kind of frenzied tap dance. He saw Harry and immediately began to eat Hermione’s family.”

Music, novels, poetry and visual art need nuanced cultural and historical understanding to be meaningful. AI is not even close to producing the kind of authentic, creative works that humans can conjure. In fact, AI cannot grasp humour, sarcasm, irony or abstract ideas. 

“Producing a song with AI requires you to feed it thousands of rhymes and sound patterns,” says Professor Srinivasan. “You don’t have the organic process that humans have to create music, where we take inspiration from a life event or something you see.” 

Before AI can achieve the kind of creativity that humans can, says Professor Srinivasan, we first need
to figure out the basics of how AI
even works. 

“No one fully understands how AI works and why it works so well,” says Professor Srinivasan. We just feed in the data and turn the crank to get an answer—it’s a wonderful but very mechanical process.”

The answer may lie in knowing more about how the brain works, he says. “Understanding the biological short-cuts that brains use to perform the same tasks as AI can give us insights that can help to create better machine learning algorithms.”