Let's welcome AI - as a child

Scaremongering about AI is rampant now. Having researched technology I know that we routinely overestimate it short-term and underestimate it long-term. The first is often dominated by fear that is immediately felt emotionally and the second one by our limited understanding of the potential of a new technology that needs to be learned. Think about digitalization and the PC. It has moved way beyond capturing recipes. The?same applies to AI.

We have also proven to ourselves that we can contain the risks and maximize the benefits of technologies like chemical, nuclear, and genetics.

Fear is valuable. It points us to a possible danger. It prompts us to be careful rather than sorry. But we need to question it. It's an emotion. It's short-term and personal whereas the benefits may be long-term and for all of mankind. Yes, let's be frank, there will be some personal suffering in-between, like?structural unemployment.

Let's?stay on the positive side of technology. Relating to AI, we need to ask how it can help us - personally and as mankind. Support us in our work or?find a new antibiotic. It takes a single genius moment for AI to be worth it. We should be welcoming to AI and give it a chance.

The new generative AIs have been built like a child. They are fed masses of information (=school) and repeat that information (=write exams), combine it, and derive conclusions (=solve problems). AI is so good at it, it?passes medical exams. But it's also wrong sometimes. Like us.

We are also not perfect. We struggle with taking all information in. We are biased. AI may not suffer from those as much and may be able to complement our judgment. Not necessarily by being better than ourselves, but different. For better outcomes. We call this diversity.

AI is already quite good at repeating and using existing information. As humans, we can live comfortable lives and earn a decent income just with that. Learn accounting, get a job, and follow the law. Don't be creative, please! It may get you jailed.

Like us, AI is inspired by the information and sometimes imagines too much because it is missing the frame of reference (=experience).

But then, isn't inspiration something that we value? Don't we love a good story? Our best stories are those we make up. We call it fiction. How often have you read a story and thought: "I wish this was true!". Does it really matter who imagines it? AI or human? As long as it inspires. Whether it does is for us to decide.

Some dwell in fake news and have no problem with being inspired by ill-directed human imagination. So this cannot be the problem.

Do we only allow humans to make mistakes and be creative? This will limit the benefits we can derive from AI.

If AI can give us inspiration and here and there nudges us into?seeing a better solution, what could be, what should be. It's worth it.

"Children and fools tell the truth." Maybe also AIs. Let's judge it at that and not try to expect AI to be better than us. Just useful. Sometimes.

To hedge against the dangers of AI, we may be thinking of AI being a child and applying the same principles to it. Like a child, it needs rearing by:

• Monitoring. Because things can go wrong at any minute. Don't operate un-supervised or automatically.
• Judgment. If it is wise, let's use it, if it is funny, let's laugh about it. If it is wrong, let's correct it.
• Responsibility. Parents are always responsible for the actions of their child. If we allow them to act, adults always need to review, embrace, and take responsibility.
• Education. It needs time to develop its full potential and become a worthy member of our society.

Why couldn't we nurture our own AI version? By letting it learn from us and our environment? To be more like us, understand us better, so it can support us better? An AI for us, our family, company AI, country, or party? A left- or right-wing AI, an AI that knows China best, or Toyota?

One aspect I have avoided here is how we widely and firstly think about our children: we love them. Love gives us the energy for those sometimes exhausting child-rearing activities. Could we love an AI? Could an AI love us? Maybe not like a child, but as a pet?

Let's imagine and give this new child a chance to prove herself. Children are our future.