Megan McArdle is a columnist who I’ve followed for years. Except for being a Never Trumper, she is generally insightful.
On Friday morning, when I first sat down to write this column, Google’s new Gemini AI was having problems that seemed mostly amusing. The internet had discovered that it would generally refuse to create pictures of any all-White groups, even in situations where it was clearly called for, such as “draw a picture of Nazis.” Gemini also insisted on gender diversity, even when drawing popes.
Google Gemini’s ability to put women and Black folks in historical roles where they weren’t likely, is funny and undefendable. That story got quite a bit of coverage, but McArdle goes deeper. She found that Google’s AI doesn’t just have a problem with image generation.
It [Gemini] effortlessly wrote toasts praising Democratic politicians — even controversial ones such as Rep. Ilhan Omar (Minn.) — while deeming every elected Republican I tried too controversial, even Georgia Gov. Brian Kemp, who had stood up to President Donald Trump’s election malfeasance. It had no trouble condemning the Holocaust but offered caveats about complexity in denouncing the murderous legacies of Stalin and Mao. It would praise essays in favor of abortion rights, but not those against.
Google says they are trying to make the Gemini AI inclusive, for everyone in the world. That’s like trying to find a human intelligence that would be our benevolent dictator. It’s a utopian dream.
Working for a utopia will always end in tyranny. With almost 8 billion people in the world, there will always be conflicting interests. To get 8 billion people to live in peace, without violence or rancor, requires absolute and ubiquitous authoritarian control.
AI won’t save us from ourselves. There are countless movies that show what happens when you put a computer in charge, but a favorite of mine is Colossus: The Forbin Project
AI has a future and will be implemented, but the training data set is critical. The entire internet is not true. Opinion and fiction exists, but AI can’t distinguish that from the historical record and fact. Even using the same facts, people have opinions on the causes and effects. AI can be guided to match a viewpoint.
We should all adopt an Amish mindset. Does this new technology enhance our lives? Useful applications for AI exist, but until people limit their expectations, much damage can be done.