Newsletter

9to5Google Log Out: It’s a shame that ‘AI’ is losing its sci-fi definition

Abner Li | Apr 4 2023 - 2:00 pm PT

9to5Google has a rebooted newsletter that highlights the biggest Google stories with added commentary and other tidbits. Sign up to get it early in your inbox, or continue reading 9to5Google Log Out below:

As a science fiction fan, artificial intelligence has always meant computers, robots, etc. that can think, with the “artificial” in the name qualifying that it’s non-human in origin. I would say that most people tangentially shared this definition as well until recent years.

In this modern era, everything is said to be powered by AI. For a time, the smart things our phones did were marketed as machine learning. To me, that felt like a fair term as it maintained my sci-fi definition of AI. Recently, I feel like companies have given up on trying to distinguish between AI and other methods.

In the case of Google, I feel like the company – as of late – has been particularly guilty of this:

9 ways we use AI in our products: High-level examples cover features in Search, Maps, Pixel, Photos, YouTube, Assistant, Gmail, Ads, and Cloud
7 ways AI is already making your Pixel more helpful: Specific examples –
- Magic Eraser
- Photo Unblur
- Real Tone
- Super Res Zoom
- Clear Calling, Call Screen
- Guided Frame (in Camera app)
- Live Translate
6 ways Google AI is helping you sleep better:
- Sleep Tracking
- Sleep Profile
- Sleep Score
- Cough and snore detection, Smart Wake
- Sleep Sensing

I’m not sure I entirely blame them either for this shortcut. Getting across these concepts to end users is not an easy task. People are vaguely aware of AI or, at the very least, associate it with something super advanced and futuristic. As such, “AI” serves as a marketing shorthand to say a company or product is invested in cutting edge technology.

Looking at what’s being given the “AI” label today, I feel that “artificially intelligent” is the more accurate meaning for that abbreviation. Artificially intelligent products and features are quite smart and might be better and faster than any humans at accomplishing a certain task (in the way that computers have infinite and instantaneous processing capabilities). However, they don’t actually think, reason, or even know (what’s real) compared to a human. For example, today’s large language models (LLMs) don’t know what’s factually real and are therefore prone to hallucinating facts.

I just regret that the term AI is no longer being reserved for what I think perfectly describes a worthy and ambitious goal of technology and computer science. But alas, we’re not going back, and everything is now “AI” to the point where what I actually think is deserving of “AI” is now called artificial general intelligence (AGI).