AI for Peace, Bombs for Children

While children die in a bombed elementary school, the US First Lady speaks of peace through AI. Over 100 children died when the US bombed a school in late February.

Just a few days later, Melania Trump stood in the UN Security Council (the organization’s most powerful body) and declared that the US stands on “the side of children around the world.”

She claimed that lasting peace will be achievable if we simply value understanding each other, and that we will be able to foster a peaceful world if we succeed in empowering our children through education and technology. The solution was spelled artificial intelligence (AI), which was claimed to be able to democratize knowledge, and thereby bring about a more peaceful world. Yet another false dichotomy was discernible here: If you are skeptical of AI, then you are for war.

It brings to mind the naive hopes of the 1990s about the internet—hopes that were soon dashed. Instead of democratization, we got platforms like Facebook—surveillance machines where the users themselves became the product. Instead of free exchange of opinions between people, we got deplatforming, censorship, and control. Researchers estimate that search engines like Google can have a dramatic impact on public elections—without leaving any traces for authorities. Platforms can also distort the information that is allowed to spread, which can have similar effects. Already fifteen years ago, Facebook experimented with its users to investigate if they could influence their users’ emotional state, which worked perfectly fine.

People have manipulated information for a long time, but the manipulation can reach new heights with AI. AI can persuade us with a skill that surpasses humans, making the risk of mass manipulation immediate. A recent Israeli investigation showed that Prime Minister Netanyahu and his allies are using armies of web bots and AI to influence the discussion on social media: approximately half of the political accounts in Israel were estimated to be fake accounts or bots. In today’s AI use, people are also afflicted by paranoid delusions. Researchers do hope to be able to use AI to cure depression, but if their technique is successful, it will also be able to induce depression.

We must do something for the sake of the children according to Melania Trump, and that something is AI. The use of AI is already spreading across more and more areas, ranging from dairy farming to the war industry. Like other tools, however, AI has no will of its own; its impact depends on how we use the tools (and against whom). Today, Israel uses AI to select targets in Gaza, and the US to select targets (like elementary schools?) in Iran.

“A computer can never be held responsible,” states a training manual from the computer company IBM in the 1970s. “Therefore, a computer should never make any management decisions.”

Before we embrace AI as the savior of us and our children, perhaps we should ask the questions that IBM already pointed to in the 1970s: who controls it, what values are programmed into it? And how do we hold someone accountable?