Seth Earley kicked off the IA Summit with a keynote about artificial intelligence.
First up is @sethearley, ceo of Earley Information Science, for There Is No Artifical Intelligence without Information Architects. https://t.co/paR921uWPy
— Anne Gibson (@perpendicularme) March 23, 2018
About 3 years ago, an analyst said “your company will not be relevant anymore.” The work that you do as an IA or content strategy won’t be needed anymore because of data, content, and information projects #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
There’s been a lot of smoke and mirrors in artificial intelligence. “How do you evolve your chat bot?” And it was a question/answer pair set… not even machine learning. Not scaleable #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
“The customer has the knowledge base” was the assumption of the chatbot producer. That’s not a safe assumption #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
There’s a lot of assumptions by AI companies that promise the AI can handle taxonomy an IA and other things, but there’s nothing to back it up #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Seth went from knowing nothing about the industry to being a member of the SWAT team for IBM’s CEO. You’ve never done what you’ve never done. People have to start somewhere. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
We have much more than we’ve ever had. Think of the things you have and see how you can reapply them. Push yourself. There are obviously some limitations – not going to be a pro athlete this year – but you have the ability to apply your skills in new ways #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Machine learning example: he uploaded a picture of himself with a beard to https://t.co/gwl6lYV17p and it guessed 56. Without the beard it guessed 74. (Also don’t use artificial intelligence to try to win an argument with your wife.) #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Seth Early to play Andy Dufresne, in The Shawshank Redemption 2 #IAS18 pic.twitter.com/EP6tXXj55M
— Richard Dalton (@mauvyrusset) March 23, 2018
Everything in our industry is moving in exponential hockey-stick growth. It’s hard to appreciate it. Smartphones were a physical impossibility 30 years ago and not even conceivable 20 years ago. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Every aspect of this growth is about information. We don’t know where we’re going but we know it’ll be based on information, data, contact, and knowledge. Information creates value when you add it to materials. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
A smart phone is sand, oil, metal, and tens of thousands of person-years of knowledge. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Architecting information speeds the “information metabolism” of the enterprise, and turns up the clock speed of value creation. All an enterprise does is manage information, even if the enterprise is a manufacturing company. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Our whole world is information and our whole digital world is information. There’s so much value to this process. People don’t really have that visceral understanding of the value of that information. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Video of the animation of DNA being replicated. We are information. #ias18 pic.twitter.com/4Y07INPFgS
— Anne Gibson (@perpendicularme) March 23, 2018
Life itself is information. How does artificial intelligence fit into this? It’s an evolution of stuff we’ve already been doing for a very long time. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Executives are seeing a lot of shininess when it comes to artificial intelligence. Articles: “What seems to be AI is really vast knowledge combined with sophisticated UX.” “The definition of AI has been stretched so it generally encompasses anything with an algorithm” #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
“When it works, it’s not AI” – spellcheck, things like that, are artificial intelligence except nobody says that it’s AI #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Search for “obsolete jobs,” get a CareerBuilder add saying you can search and apply for obsolete jobs on their site. Oh, AI. #ias18
— Dylan Wilbanks, now at @dylanw@xoxo.zone (@dylanw) March 23, 2018
Enterprise AI: two ends of the spectrum: one end is heavy-duty algorithms, heavy mathematics. The other end is the work we do as IAs that someone is calling AI. Text analytics, a long-time staple of content management, is now called “AI”. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
We have a lot of things coming together and there’s going to be massive changes to the job market — but that’s always been true. Technology has always caused job disruption throughout history. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Cognitive computing is a better way to interact with machines. That’s what we do, that’s what we’ve been doing. Forrester calls it “Knowledge Management’s Grand Makeover”. It’s a rebranding #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
We’re trying to reduce the cognitive load. Chatbots, intelligent virtual assistants, conversational search… they’re retrieval mechanisms on an information retrieval system. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
What kind of knowledge base do you have? How about search interaction, information architecture, user experience, enabling technology? That determines whether you’re at the “basic search engine” or “intelligent assistant” end of the continuum #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
When you search for something you don’t want a 100 results, you want an answer. @sethearley #ias18
— Carrie Hane (she/her) (@carriehd) March 23, 2018
We build on the foundational systems of the past. When we train a search engine, we train it with the knowledge we already have #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
The data that went into Watson was originally data from Lotus Knowledge Server. For all the systems, if you give it the wrong content, performance degrades rapidly. You can’t just point a system to everything, you have to contextualize. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
How IBM Watson was trained…a bunch of fancy words for "content." We can be involved in many levels. @sethearley #IAS18 pic.twitter.com/LpGG4ipuqf
— IAC – information architecture conference #IAC24 (@theiaconf) March 23, 2018
We have tons of different formats of data, tons of data, there’s a lot of phrasing and terminology that’s in our companies today, and an AI can’t understand it without understanding the information architecture and content #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
You can put a bot front-end on search. Search is a conversation. Search terms are short, ambiguous, and an approximation of what a user needs. Conversational search can help with that #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Faceted search can be reflected as a conversation, where the bot asks the right questions to limit to the facets, and prompt the user to follow the foundational architecture of the data #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
When people have conversational interactions, they treat the interaction differently. “Good morning [bot] how are you today?” You don’t walk up to people and yell “[search term]”. Conversation helps to lead to intent #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
We have ontologies, relationships, very deep, very nuanced, and then we need to apply that to a content. Chatbots are a channel. They’re evolving. There’s no knowledge engineering going on in these things #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
How humans can be involved at every stage of AI "training." @sethearley #IAS18 pic.twitter.com/tRNtTFvbIQ
— IAC – information architecture conference #IAC24 (@theiaconf) March 23, 2018
You need knowledge engineers, integration engineers, supervisors… oh hey and they do the same things as information architects. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Information architecture and information architects are essential to the process. When orgs think these are new skills and new groups they’re wasting resources #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Componentization is where we think about repurposing our data, then apply it to lots of scenarios. (Content strategy heavy here!) Take existing assets, deconstruct them to components, then reuse them in lots of contexts. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
When you start looking at building them out, think in two dimensions. The task complexity and the domain complexity. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Don’t start in the top right corner #ias18 pic.twitter.com/Gvd9DkQwdZ
— Anne Gibson (@perpendicularme) March 23, 2018
If the assumption is that your organization or the industry or the technology is that you’re going to be pushed out, rebrand into a data scientist. IAs and data scientists are very similar #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Seth is showing parts of a video by Corning called “a day made of glass” where everything is a screen. Our environment will have info contextually aware of our needs, wants, tasks, objectives. That will be surfaced as we go about our day #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Changes in material science, the internet of things, all of those allow for this — as long as there’s information architecture. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
There will be AI and machine learning as part of the future, but it will also be mostly user experience. Humans will be doing that work. #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
“There’s no AI without IA” – article
“The problem with AI” – article #IAS18— Anne Gibson (@perpendicularme) March 23, 2018
Information architects ARE data scientists. That knowledge will be key to the content that powers AI. Seth Earley’s keynote blowing my mind. #ias18 #ArtificialInteligence #UserExperience #ContentStrategy
— Susan Westwater (@SJW75) March 23, 2018
Have you ever said no to what a client wants to do? Our director of operations says no all the time. You have to say no. You have to ground people in reality. You’re better off losing an opportunity than not being successful. You absolutely have to say no #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Let them come back when they fail, because they will fail if they don’t resource appropriately #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018
Let them come back when they fail, because they will fail if they don’t resource appropriately #IAS18
— Anne Gibson (@perpendicularme) March 23, 2018