Slow Down
Once again, the technology curve is far ahead of most citizens, businesses, and governments’ abilities to keep up. Nothing shows that more than generative AI being shot like a cannon right onto the world stage and world’s screens, with advances from the likes of OpenAI, Midjourney, Bard….
This is no race to the moon — it’s a race to no-one-knows-where, and the stakes are high and the implications enormous. And it’s happening at breakneck speed, with no ground rules or regulations. Imagine Formula 1 competing this way! Now Microsoft tells us ChatGPT will be added to the Office suite… Whooooaaa. Hold on a minute.
Experimentation is fine. But throwing generative technology that has so many unanswered questions into software that is an enterprise standard makes me shudder. So it was a welcome relief this morning when I awakened to an email from my colleague David Worlock in the UK about an open-letter to pause giant AI experiments:
Here is the letter, published here this morning and signed by Elon Musk, Steve Wozniak, and over 1,000 scientists, politicians, and industry leaders, calling for a moratorium on further AI development until some of the risks have been sorted out. I’m not sure this will have much effect on anybody, and The Verge reports that some of the signatures were not inserted by their owners! However, it does underline issues of regulation and control that are going to be very much a part of the public debate in the coming year. All best wishes, David
God let’s hope we get some public debate going and the regulators pay attention. Let’s hope this call to action gets some traction. I’m still baffled by (and have spoken and written about) why my local broadcast news is regulated in this day and age, but Facebook goes unfettered.
The issues are immense. What is fair use? What is the provenance of the data these technologies are using? Is it complete? Is there bias, and if so, what kind? How old is the content? Can I reliably make decisions using it? What decisions? Life-and-death ones? Let-people-out-of-prison ones?
What are the implications on how it can improve teaching and outcomes? What applications is it best suited for, and what happens to my IP when I use it in my enterprise? Where does it go when it’s in commercial software — are there new Ts and Cs? What security and privacy implications am I unleashing? And what happens to any outputs that I commercialize? What happens when my outputs become new inputs?
I am dizzy.
While some of the experimentation will yield good answers, embedding this stuff in mainstream software is way, way, way too premature, and Microsoft should know better. It’s tech’s (remember those sheep?) obsession with being first that has these companies behaving wildly, and it’s terrifying.
I am all for new technology — all for innovation. It’s when the risks are too high and there isn’t time to figure out the unknowns that it’s worrisome. We’ll never figure it all out, but warp-speed 10? Why? Why do we regulate seat belts and — yes, this really happened! — whether I can add my own tea bag to scalding hot water in an English convention call, but OpenAI gets to implicate itself into my computer whether I want it or not? Please heed the memo — regardless of whether Messrs. Wozniak and Musk really signed!
Let’s take a few minutes to find out. Please make sure the regulators pay attention. And the industry associations — is anyone out there listening? The GitHub lawsuit is telling; Getty putting a stake in the ground is, too. Once again, private enterprise will do the job of fighting back, but it will take more than lawsuits doing the heavy lifting to get this thing right.
And don’t miss our series on the important questions and fact-based answers. We are on the case, whether this week’s analysis looking at enabling technologies impacting our industry or Hugh Logue’s latest piece on the EU’s stance on generative AI and how it’ll likely unfold. (These pieces are for members only — it’s time to become one if you’re not.)
There is simply too much going on and too much noise to rely solely on a search engine. And relying on Bard and ChatGPT to keep telling us about themselves? Oh, my….