Believe in Smart Software? Sure, Why Not?

January 12, 2024

green-dino_thumb_thumb_thumbThis essay is the work of a dumb dinobaby. No smart software required.

Predictions are slippery fish. Grab one, a foot long Lake Michigan beastie. Now hold on. Wow, that looked easy. Predictions are similar. But slippery fish can get away or flop around and make those in the boat look silly. I thought about fish and predictions when I read “What AI will Never Be Able To Do.” The essay is a replay of an answer from an AI or smart software system.

My initial reaction was that someone came up with a blog post that required Google Bard and what seems to be minimal effort to create. I am thinking about how a high school student might rely on ChatGPT to write an essay about a current event or a how-to essay. I reread the write up and formulated several observations. The table below presents the “prediction” and my comment about that statement. I end the essay with a general comment about smart software.

image

The presentation of word salad reassurances underscores a fundamental problem of smart software. The system can be tuned to reassure. At the same time, the companies operating the software can steer, shape, and weaponize the information presented. Those without the intellectual equipment to research and reason about outputs are likely to accept the answers. The deterioration of education in the US and other countries virtually guarantees that smart software will replace critical thinking for many people.

Don’t believe me. Ask one of the engineers working on next generation smart software. Just don’t ask the systems or the people who use another outfit’s software to do the thinking.

Stephen E Arnold, January 12, 2024

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta