I understand that this is merely a developer beta, and if queried by anyone at Apple, my first impressions of Visual Intelligence should solely reflect its future possibilities. Nevertheless, I find great excitement in trying it out and envisioning what’s ahead.
The current functionalities of Apple Intelligence are quite basic. Mostly, users can choose between using ChatGPT to provide descriptions of what they see or conducting a Google search …
The Core Functions of Apple Intelligence
Based on my observations, Apple Intelligence performs a few key functions:
- Summarizing text present in the visual scene
- Providing business information from Apple Maps
- Identifying dates and times in text, automatically opening Calendar for those entries
However, when it comes to recognizing objects, the feature primarily relies on ChatGPT and Google. From my limited experience, Google stands out as the more impressive option.
Image Recognition: Google and ChatGPT
I typically conduct thorough research on my significant purchases before making a decision. I read multiple reviews, seek recommendations from friends, and extensively use Google for details.
Therefore, when I spot something in action and want to research it, I need precise details: simply stating that it’s a compact bean-to-cup coffee machine is not sufficient.
So, how does Visual Intelligence v0.1 perform? I tested it on several items I had at home.
First, I tried it on my Sage Combi Wave 3-in-1, which combines a microwave, oven, and air fryer. Known as Breville in the US, Visual Intelligence quickly identified it (first match, top right). It also suggested a competing product, but the remaining top results were accurate.
Next, I tested it with my Sage Barista Touch coffee maker. Here, Visual Intelligence was less certain, suggesting either that model or the more expensive Oracle Touch from the same line. Nevertheless, it was easy to determine which option was correct. (It also presented a third model, which turned out to be an ad with a mislabeled picture.)
When it came to my Aarke Carbonator II sparkling water maker? No issues there.
It had some challenges recognizing a Reidel O Wine Tumbler, but that’s understandable given the numerous imitations available. The program primarily showcased the original and a few direct replicas.
When I moved to the living room, it successfully identified this as an Eames Lounge Chair, and considering it’s a precise replica, I can excuse the lack of differentiation between it and the authentic item.
I tried identifying some original artwork by lesser-known (yet recognizable) artists, but it didn’t succeed. When it came to a print, however, it had no trouble recognizing it.
As for my Naim Mu-so, it mistakenly assumed it was the second-generation model rather than the first. Given their identical appearances and the abundance of references online concerning that model, this is entirely justifiable.
I further attempted to identify several other tech gadgets around my home:
- Original HomePod
- Echo Dot
- Philips Hue Switch
- Netatmo Smart Doorbell
Visual Intelligence successfully identified all of these.
A Promising Start with Much More Ahead
My initial experience with AI-powered image recognition technology was during my time with the Ray-Ban Meta glasses. However, that required careful phrasing and follow-up questions for optimal performance.
Visual Intelligence already shows improvement in this regard, which is what excites me most about its evolution. With numerous companies vying for advancements in this sector, I anticipate rapid progress soon.
Of course, there will be mistakes along the way, as with all AI technologies, but this sort of technology learns and improves with each interaction.
It won’t be long before we can point an iPhone at a random product in a store or on the street and instantly receive an average star rating along with a comprehensive AI-generated review summary. Typically, Apple announces that they can’t wait to see how we use their new technology; in this instance, the anticipation seems to be mutual.
If you have access (which took me a couple of weeks to obtain), I would love to hear your first impressions of Visual Intelligence. Please share your thoughts in the comments section, and check out Fernando’s video below.
Photo: DMN
FTC: We use income earning auto affiliate links. More.
