Will AI enslave us?

Artificial intelligence is a tool to discover patterns in very large datasets. ChatGPT, for example, uses a form of AI that combs through hundreds of billions of documents and images to find plausible ways to respond to a given question.

Other forms of AI discover patterns in videos and even sound recordings. Recent results have been astonishing.

But to answer the underlying question, no: AI will not enslave us.

ChatGPT derives its wisdom from documents written by people. It is merely a mirror of how we communicate, not a malevolent force that can reduce our civilization to ruin.

Like any tool, AI can be used as a weapon, and this is something I worry about. Bad actors already use AI to generate fake news. With the most recent advancements, they can now create fake voice recordings, fake images and fake videos that are indistinguishable from reality. Such acts will lead to new forms of cybercrime and new threats to our democracy.

AI also allows authoritarian governments to spy on people at levels never experienced. And AI will lead to more potent cyber-attacks on our infrastructure, our corporations and our democracy. These are all threats I take seriously, and about which we should all worry.

But in all this, we should not forget why we are pursuing development of artificial intelligence in the first place.

AI has already saved countless lives by helping doctors to diagnose deadly diseases such as cancer.

As I write this, a plethora of driverless cars are operating in my neighborhood in San Francisco, bringing an unprecedented level of safety and access to transportation to us all.

AI also has become an indispensable tool for creatives, professionals who generate content for marketing, education and entertainment, and even software engineers.

Udacity now provides personalized AI mentors to more than 3 million students in the Arab-speaking world and Uzbekistan. Cresta provides AI coaches to call center agents. A recent study by researchers from MIT and Stanford found a 14 percent improvement in productivity.  And AI has long been used by companies such as Google to find the information you are seeking.

In the present debate, we are missing the voice of reason. Some of our leaders link AI to nuclear and pandemic apocalypses. This is not the future I see. I see a technology that will make all of us better people.

Think how much of your daily work is mind-numbing, repetitive and unenjoyable. You will soon have your personal AI assistant, to whom you can hand over your menial tasks, freeing up your mind and your time. The assistant will be entirely under your control. And all children will have personalized AI tutors. 

We should all welcome a broad debate about the pros and cons of AI. But let’s not forget that AI is a tool used by people, which derives all of its signs of intelligence from things that other people have written. AI is not a living being that has evolved to survive, but a tool developed by us.

Like a kitchen knife, AI can be used as a tool or as a weapon. Let’s make AI serve for everyone’s benefit, and let’s work hard to prevent abuses. 

Sebastian Thrun is an adjunct professor at Stanford and a pioneer in the field of AI. He co-founded Google X, Waymo and Udacity.