Ai DayAI SecurityAI & DataBlogSecurity
Blog • • 4 minute read

AI Flavors…

We're continually bombarded with various types, flavors and descriptors of what AI is (and isn't). Therefore, to make things a little simpler (and to follow up on the on-stage descriptions I gave of the AI variants), this is a short description of what three of the main AI options would do with (and for) you if you'd ask…

Given we've gone from narrow AI to general AI, then generative, natural language processing, large language models, and now agentic AI seemingly taking over all the headlines let's take a moment to baseline what each is (and isn't) using a simple (and hopefully) effective example:

 

The scene:

You're at home, it's the morning before your trip, and you're finishing up the packing…

 

1: "Hey Siri, what's the weather in {SELECT city.name}"

Natural language processing (NLP) and machine learning kicks into action, recognizing your question (thanks to the language component OF NLP) it queries and then pulls structured data from weather sources, possibly a few different sources if we're using an efficient narrow AI task-based algorithm to consolidate sources into one good answer and then presents you with the results…

"The weather in Lerwick will be warm during the day, and then a little cooler at night"

  

2: "Hey Siri, do you think I should bring a sweatshirt with me for the trip this week to Shetland/Lerwick?"

Now we have a variable thrown into the mix that adds context to the whole question. This is where generative AI comes into play where the data (weather) and context of the request (temperature, clothing, time of year, and if wearing a Fair Isle jumper is welcomed or a faux pas for foreigners/visitors). Once it's done the analysis it generates an informed suggestion based on the knowledge sources as opposed to simply repeating known facts like a parrot.

"Chris, I think you should bring your sweatshirt with you as it will blend in, and be warmer for you during the evenings"

 

NOTE: IF I'd given our generative AI access to my photos as a data source, and maybe my shopping habits it might also have pointed out that the 1990's Shetland knitted sweater I have that's green, and brown might NOT be suitable based on the current fashion etc.

  

3: "Hey Siri, can you please put a packing list and itinerary together for me for this upcoming trip to Lerwick"

And here comes the agentic experience, not just doing what you asked, but also anticipating any possible complications and autonomously making decisions for you within the context of the end goal (to get you to Shetland in as organized a way as possible) Let's work through the possible steps, permutations, and things that a much more agentic focused system could do:

- Checks the weather, works out suitable clothing choices AND quantities

- Checks your calendar, notices a mix of business meetings and sightseeing/spare time, re-checks clothing for right balance

- Assesses your flight information, sees you have a tight connection in Boston, adjusts for carry-on capacity only

- Doubles up a few items (you're now wearing your WWT sweatshirt for meetings AND one day on the hills…)

- Checked the B&B in Brae for washer/dryer facilities

- Spins up an agent to compose and send an email to double check they can allow you to wash your things

- Knows you must get from the mainland to Yell, and maybe Unst so makes sure you have a rental car waiting for you

- Realizes that the islands are not connected, spins up another agent to plan the inter-island ferries

- Checks you into the flight (another agent spun up) cross references the seating and meal preferences to ensure all correct

- Collaborates with your calendars, fills in the necessary points that you must pay attention to

- Notifies work on the meetings, and friends/family on your free time.

- Double checks names against folks from your contacts to ensure communications right/correct (context etc.)

- Wraps all of this up nicely, tidily, and makes sure you have the right PPTX files for the meetings

"Chris, all done, your hotels sorted, the packing list is ready on the printer, and the meetings and friends know you're coming"

 

NOTE: This is agentic AI, it knows enough about you, OR knows where the data is about you to make informed decisions on your behalf. It has access to the necessary data sources to both understand context AND adapt based on known variables (packing case size vs. carry on vs. clothing capacity etc.) It knows what you typically spend so will allow for space for gifts on the way back (has access to accounts) basically consider it "your" shadow learning as it goes along, and as you create (and feed it) more data points for it to learn from.

 

PS: Before anyone goes and asks Siri to plan out their next holiday PLEASE know that our colleagues over at Apple have not  (thankfully) unleashed agentic AI onto that platform yet, so please just keep the questions to nice, simple, focused weather related ones for now.

'all for now

Chris