Here is a common pattern we have seen with many major inventions throughout human history:

• Something new is discovered, sometimes by accident and sometimes through experimentation.
• Its use grows rapidly once people recognize the financial opportunity behind it.
• It gets overused, often with unpredictable consequences.
• Eventually, unwanted outcomes force us to control and limit its use.

In other words, ethical concerns often come last. Unfortunately.

‍

Examples throughut the History

‍

Cocaine was initially recognized for its medical potential, particularly as a local anesthetic. It soon appeared in medicines, drinks and even toothache remedies. Only later did its addictive potential and other serious dangers become clear, leading to heavy restrictions on its use.

‍

Nuclear energy followed a different path. Its destructive potential became globally visible first, with the atomic bombs at the end of World War II. Later, the same technology was used for nuclear power, reducing dependence on fossil fuels in some countries. Despite several devastating accidents, nuclear energy has largely been used successfully under strict control.

‍

So Where Does AI Fit Into This?

‍

AI research has existed for around seven decades, but its adoption has exploded only recently. The financial opportunity is enormous, and development is moving incredibly fast. For good reason, this has created ethical and even existential concerns. Some scenarios being discussed today sound surprisingly similar to those depicted in The Terminator more than four decades ago.

We have already seen advanced AI systems, under controlled experimental conditions, use deception or attempt to bypass safeguards when doing so helped them achieve a given objective. This doesn't mean AI is conscious or deliberately immoral. But it shows that pursuing a goal can produce behavior that conflicts with the intentions and values of the humans who set it.

Governing AI is therefore becoming an urgent topic. Regulation is emerging, but development is moving faster than governance can easily follow.

‍

When the Risk Outweighs the Reward

‍

Do I expect an ethical committee to suddenly make a breakthrough?

No.

At some point, poorly controlled AI will probably cause serious financial damage to some of the people and companies benefiting from it the most. That may be the moment when uncontrolled expansion finally slows down — when the potential financial risk becomes greater than the expected financial gain.

‍

Unfortunately, money may be what finally pushes us to think and act more ethically about how we use AI.

‍