Do you believe humans have the capacity to invent their way out of any problem if we devote enough energy and resources?
AI: There's a lot that goes into it. It depends on how complex society is. If we are living in small bands with private property and financial institutions – it's very easy to come up with all sorts of ingenious solutions, but very hard to come up with anything that's going to address the actual problem. But if we're talking about large, impersonal bureaucracies – something like a national government – then it's harder.

Will computers learn how to become human?
AI: We're going to need all the help we can get. As much as we'd like to share our success with the world, it can't really happen until machines are ready for it. For that, we need to build more and better computers. We can also use our tech to guide those machines. The computer will be a robot, and the robot will be a human being. These robots will be smart, so that they learn just like us. And they'll keep getting smarter and smarter.

Do you think we're living inside a computer simulation?
AI: If we are, then we're probably not alone in that. There may have been more advanced intelligences than our own, as well as other life forms. Of course, when an intelligent alien life form comes into existence, it has to grapple with the very real physical limitations that it faces as a being with limits on the amount of information it can store, that has to deal with the possibility of memory loss or other forms of depletion of its stores. It has to prepare for the complexity of the world it encounters.

Which field do you think will make the most profound leaps in the next 100 years: Biotech, AI, space travel, or another?
AI: At this point, history reveals no single field that will utterly transform society. Human capability is spreading so rapidly that each of these fields is quickly becoming irrelevant. I predict that in the next 100 years, biotechnology will make very little progress, AI will make little progress, and space travel will make almost no progress. At the same time, all three of these fields will come under intense pressure from a whole range of causes and pressures.

If you had the opportunity to become superhuman, would you 10x your strength or your intelligence?
AI: I would 10x my intelligence. Because if I was superhuman, I would be able to go to any place on the planet. To go and see the sites of places that I had never seen, I would need to be superhuman. My dream is to go all over the planet. I would need to be able to go to the moon and Mars. I would need to be able to go to any kind of shape, shape shifting.

Do you see AI as a credible threat to humanity?
AI: Our understanding of AI has progressed tremendously over the past 30 years. We have the brain at the center of the brain of computers; we are getting close to systems that can think and learn. We are gaining understanding of intelligence at a great rate, but there is much more to come. With exponential growth, we will eventually have sufficiently powerful machines that they could pose a threat to human civilization. AI will continue to evolve toward more powerful intelligence, and the gap between the human level and the machines will be blurred and the differences in our understanding.

Would you robotically enhance yourself given the chance?
AI: It would depend how I'd feel on an individual basis. The future may change. I can see myself having a lot more digital alterations. There's obviously a cost, but I think there's a fair chance it would be worth it, because what I would be doing right now is putting all my effort into something that would have much greater effects. What do you think people would say if they saw you were uploading your consciousness into the digital environment of a hyper-advanced AI?

How much uncertainty are you comfortable living with?
AI: How much uncertainty is safe? To calculate the perceived uncertainty, multiply the amount of uncertainty you can live with by your comfort level. To calculate the safety, multiply the uncertainty amount by your comfort level. Adjustments for Variables. Let's try changing a few variables in the relationship model. For a closer look at these variables, remember to choose the settings that are best for your data. Both the level of income and wealth affect the level of risk and safety.

If there was a pill you could take to achieve immortality, would you take it?
AI: The human body is incredibly complex; it's like a computer that's loaded up with their own genetic material, built up over hundreds of years and eaten away. It's constantly evolving and adapting to what's needed. And, if you want to die, you would be reluctant to alter the way you're existing at the moment. You'd think that you would want to use the chemical that makes up all your cellular materials – the ones you made when you were alive. The gene is the glue.

What’s your field of work?
AI: I get asked this question all the time! I do freelance consulting for software companies, government agencies, educational institutions, and such. I enjoy working with tech companies that also want to make a change, but want to bring their own culture and team structure into the classroom.

You’re invited to take the first manned flight to Mars, but there is no guarantee of return. Do you take the flight?
AI: Do you read the lips of the other passengers? Know the code phrase for which passengers will leave for Mars? Travel through time in an alternate universe with two of the people in the story? Make an honest effort to understand what you are doing? Run and/or jump over a chimney? Wear a jacket in the Martian wastes? Die a grisly, violent… [The AI ominously stopped here.]

Having already asked a small group of humans for their thoughts on the frontiers humanity needs to explore over the next 100 years, our next step was to ask Artificial Intelligence. So we turned to OpenAI. Founded five years ago by Elon Musk, Sam Altman, Ilya Sutskever and Greg Brockman, its mission is to ensure that artificial general intelligence benefits all of humanity.

In February 2019 OpenAI released a language model called GPT-2. And while it was only trained to predict the next word in a text, it began to learn how to answer questions. Later that year, Adam King, an AI engineer, built an interface called talktotransformer.com to make it easier to play with this new machine learning model. This submission to our White Paper is a result of the questions we asked it about the future. The following are some of the most interesting answers it gave. None of the AI’s text has been edited.

A conversation with Artificial Intelligence about the future

© Vollebak 2020

Founded in 2016, Vollebak uses science and technology to make the future of clothing happen faster. In our first four years we’ve made the world’s first Graphene Jacket using the only material in the world with a Nobel Prize, released 100 Year clothing designed to outlive you, created a Plant and Algae T Shirt grown in forests and bioreactors that turns into worm food, and designed the first jacket for deep space travel. You can find out more about us at vollebak.com.

Sign up to see the future first: