The 2 big AI stories of the last week: Google and Europe
Also, I'll be at BrXnd, the AI for marketing conference tomorrow -- will you?
I had a different newsletter planned for today, but that one will have to wait, because these two stories are so big:
Generative AI will be everywhere in Google Workspace
Last week at their annual I/O conference, Google announced a slate of generative AI products in their release pipeline — features you and I, just a half a year ago, probably would have considered wildly speculative and possibly requiring of sorcery. Mashable and the Verge have some feature lists, including:
Gmail will draft full emails for you.
Google searches will soon produce AI answers.
You’ll be able to generative images from text directly inside of Google Slides.
Very few of these tools will be the first of their kind on the market. For instance, the startup Flowrite already drafts emails; Bing already has incorporated AI answers into its search engine interface, and the slide deck startup Tome already has text-to-image generation.
But from a management and workplace productivity perspective, it’s one thing to ask adopt an entirely new tool. It’s a much easier proposition to adopt a new feature inside a tool (like Gmail or Google Slides) that you already use. So Google’s announcements may massively speed up the timeline on which generative AI tools will be adopted by the masses.
Note: Microsoft has already promised similar features inside of Office 365. When I first watched their Microsoft 365 Copilot demo all the way back in the ancient days of March 2023, I uttered the shocking words “I might… switch from Google to… Microsoft??”
Here’s a 10 min demo video. I think it’s safe to say that most of the productivity features shown here will also be on the Google roadmap even if they didn’t promise it at last week’s I/O:
Of if you’re in a rush, here’s the 1.5-min version.
Probably timed to try to steal some of the Google I/O press, Microsoft had an announcement of their own last week — they just announced the expansion of their Office Copilot beta from 20 customers to 600, and with some new features added. If you’re using Microsoft Teams, Microsoft Office Suite, etc., my guess is that some of these features will roll out to you by the end of the year.
Meanwhile, the EU might ban most generative AI models
I haven’t yet laid out my views on AI risk or AI policy in this newsletter. Overall, while in the work of AI Impact Lab I’m mostly focused on the good that AI can do, I also think that advances in generative AI do pose massive risks to the world, and I’m pro-regulation. At least, pro-sensible-regulation.
Whether the regulations the EU Parliament passed out of committee today are sensible is a good question. If they became law, they would require generative AI models like ChatGPT to be tested and licensed before being made available in the EU — which at a high level seems like a good idea to me. But here’s one take arguing that these rules would have the large side effects of cementing the power of big tech companies at the expense of startups, neutering the open source AI community, and potentially incentivizing unsafe AI over safe AI.
I’d love to hear in the comments from any AI For Good readers who are following these rules about your take.
Anyone else going to BrXnd?
I have to cut this newsletter short because I’m about to board a plane to NYC for BrXnd, the inaugural Marketing X AI conference. If you’re reading this and will be there, drop me a note and let’s try to meet up!