Google I/O 2023 kicked off and the proceedings could be summed up in two words: Generative AI. Alphabet CEO Sundar Pichai and Google Cloud CEO Thomas Kurian walked through Google’s new models, how it’ll incorporate generative AI across the product line and advance business transformation.
Here’s a look at the hits and misses from Google’s strategy session on generative AI. Keep in mind, that this launch follows a rocky debut of Bard, which was upstaged by Microsoft’s ChatGPT rollout. In some areas, Google shined with its generative AI plans. In others, Google is playing catchup to the combination of Microsoft and OpenAI.
Hits
-
The Gmail feature of AI writing a contextual-based email is pretty good. You can have AI write emails based on context with short one-liner responses or detailed. The Help Me Write feature should boost productivity.
-
Google is known for its map capabilities and analysis of traffic patterns. Now with AI integration, someone can integrate traffic, weather, air quality, etc. when mapping out a route at different times and planning a trip ahead. The immersive view looks intriguing. When it comes to mapping Google owns it.
-
AI for image editing is very good. Available until now on Pixel phones only, it moves to Google Photos now. In addition to the Magic Eraser, the Magic Editor can allow for contextual editing easier without the need for expensive tools from Adobe. The additional integration with Adobe Firefly for graphic workers seems compelling.
-
The LLM model called Gemini when ready will be available in multiple sizes based on needs instead of boiling the ocean down like ChatGPT-4, etc.
-
I really like the option of identifying synthetic content with a combination of water marking and metadata embedding. This by far is the best responsible AI theme I have seen. As I repeated many times, fake content, especially during election cycles, can do much harm. This is a good start.
-
Generative AI integration with Google Workspace products such as Sheets, Docs, Gmail and Slides are all very good.
-
The AI-enabled "real-time" cognitive search to mix both Bard AI synthesized content along with regular Google search is brilliant. Combining those results with Google Shop can be good for websites and e-commerce enterprises. These features can be a differentiator for Google while Microsoft is struggling with the market size.
Misses
-
The variations of Codey, code generation tools offered by Google are still limited. CoPilot has been around for more than a year now and its code generation has been improving tremendously.
-
Microsoft has already started moving towards an industry-specific solution set using AI, but Google seems to be missing many steps in that direction.
-
While there were compelling releases, there was no mention of enterprise-specific integration with ERP, HR, CRM, and other systems. It is going to be hard for Google to gain traction without a broad integration strategy. Google did note a Salesforce integration.
-
While Kurian threw big numbers out there about AI-optimized infrastructure, AWS has the lead for enterprises that want to use variable sizing. Not sure if Google can catch up on that. The Google Cloud offering is extremely limited and didn't come across as compelling.
