AI will inspire sci-fi writers, not replace them

An expert on artificial intelligence and science fiction shares his thoughts on how AI can be used for innovation and what generative AI means for authors.

What attracted you to science fiction?
Science fiction (sci-fi) is a very appealing genre because it explores diverse ideas—sometimes scary and big ideas. But it involves unique thinking. So for writers, sci-fi is like an experiment.

How do you use sci-fi as a means to explore aspects of humanity and society in your research?
Sci-fi is not a source of ideas for me; it empowers me to challenge certain things. So, a seemingly impossible idea can become a reality.

Do you think that science fiction is now almost being outpaced by scientific fact?
AI is like science facts at a higher speed. Sci-fi fiction might still be ahead, but nowadays, writers are trying to write more science-fact-based ideas.

Many sci-fi writers have very different backgrounds in professions such as science and technology, the humanities, medicine, or academia. So they have an abundance of knowledge, but now they are trying to explore more unique ideas.

How does sci-fi feed off science, fact, and innovation?
Sci-fi should motivate the future—both people and ideas. So now, we are trying to study sci-fi prototyping. That's the challenge of creating a future vision for sci-fi writers and other individuals. Since sci-fi is not only about the future, it is more innovative.

How does it flow into the current state of innovation?
Sci-fi is empowering research. Despite questioning their existence, sci-fi writers are making us lean towards them because these strange ideas may be very important for society.

Does sci-fi help us test certain developments by looking further ahead to imagine them?
Simulations are important in sci-fi because researchers sometimes think that technology is very good for everyone, but this writing genre finds problems with certain technologies.

This point is very important for leading sci-fi, and that's why several Japanese companies are now focusing on it.

When it comes to AI, do you think that we feel more comfortable talking to somebody or engaging with someone that we think has human characteristics? Why can't we trust a robot?
Human-like agents are very useful and might be more attractive to people.

This doesn't mean that AI needs to mimic humans; but it needs to interact better. For example, some kind of human-like AI means you are helping people communicate better with society. I think it’s very important that our society can accept more “non-human-like” ideas. This means we make more allowances for diversity.

What about the demand and concern for AI to be contained and have careful safeguards around its future development? Would you like to see strict controls?
Actually, I think it's impossible to strictly control AI. For cases such as nuclear power or bio-tech, we need controls. However, if AI just runs on computers, such as for graphics or on personal computers, we cannot control it. But we can educate ourselves on how to use it better. For example, AI is very good at persuading people, which means it can deceive people. But if we are smarter and more aware, we can realize that this is just a machine trying to attract us.

So you're saying that regulation is sort of pointless? We need to educate, rather than regulate to try and control it.
We need to control AI ourselves, but not through rules or strict regulations. All humans need to be responsible so that we create a precedent and shape the future.

What's the responsibility and role of a technology company such as Huawei when we're developing these new technologies? Are you saying that we shouldn't have any regulations?
I think that we need regulation of companies to give warnings to people. While some technology might deceive people for dark purposes, at the same time, some people have a “right” to be deceived.
The important point is that we need to make a statement about the risks of AI—not just physical risks, but also psychological ones. You might be deceived by AI, so if you're using it or might be relying on it too much, this is important to recognize. Therefore, every industry and company should be required to be transparent and open about the technology they are using.

With AI becoming more and more skilled at replicating human writing, will AI be writing sci-fi in the future? Will we even need human authors?
I think it might be possible to generate sci-fi by AI. But in general, we want literature that is written by humans, because sci-fi writing is a form of human-to-human communication.

The fundamentals of what an author is thinking and what kind of story they are imagining cannot be replaced by generative AI. AI can help, but it doesn’t replace authors.

Einstein said that imagination is more important than knowledge. Do you agree with that? 
I totally agree with Einstein. It is even more important than before because we can share knowledge on the internet and AI even can answer its own small queries.

But imagination is the motivation for humans to thrive in a new direction. What kind of direction will be possible and where do we want to go? That question is more important to research and for the development of society.