Tag Archives: Chatbot

Privacy: Exposing AI Chat Plugin Access to User’s Conversation History.

There are many benefits for allowing third-party plugin access to a user’s Chat history. For example, an OpenAPI, ChatGPT Plugin could periodically troll through a user’s chat history and proactively follow up on a conversation thread that appears to still be open-ended. Or, periodically, the Chat Plugin could aggregate the chat history into subjects by “smart tagging” conversations and then ask the user if they want to talk about the Manchester United, football game last night. Note, in the case of OpenAPI ChatGPT, it has “Limited knowledge of world and events after 2021.” Also, note presently OpenAPI ChatGPT API or the ChatGPT plugin has not exposed the user’s chat history.

3rd Party, Security Permissions for OpenAI, ChatGPT API, and Plugins

Just like authenticating 3rd party apps with your Google credentials, allowing the app to access Google user’s data, this level of authentication should be presented to the user, i.e., “Would you like to allow XYZ ChatGPT Plugin access to your Chat History?” I’m sure there are many other security questions that could be presented to the user BEFORE they authenticate the ChatGPT plugin, such as access to personal data. For example, if the AI Chat application has access to the user’s Google Calendar and “recognizes” the user is taking a business trip next week, the Chat app can proactively ping the user a reminder to pack for warm weather, in contrast to the user’s local weather.

Grass Roots, Industry Standards Body: Defining All Aspects of AI Chat Implementations

We don’t need another big tech mogul marching up to Washington to try and scare a committee of lawmakers into the benefits of defining and enforcing legal standardization, whatever that might be for some and not for others. One of the items that was suggested is capping the sizes of AI models with oversight for exceptions. This could cripple the AI Chat evolution.

Just like we’ve had an industry standards body on the OAuth definition for implementation, another cross-industry standards body can be formed to help define all aspects of an AI Chat Implementation, technology agnostic, to help put aside the proprietary nature.

In terms of industry standards for artificial intelligence, Chat standards, permissions for the chat app, and 3rd party plugins should be high on the list of items to invoke standards.

Extensions to AI Chat – Tools in Their Hands

Far more important than the size of the AI Chat Model may be the tools or integrations to the AI Chat that should be regulated/reviewed for implementation. The knowledge base of the Chat Model may be far less impactful than what you can do with that knowledge. Just like we see in many software products, they have an ecosystem of plugins that can be integrated into the main software product, such as within JIRA or Azure DevOps marketplaces. With relatively simple implementation, some plugins may be restricted for implementation. Many AI Chat applications’ extensibility requires manual coding to integrate APIs/Tools; however, assigned API keys can solve the same issue to limit the distribution of some AI Chat tools.

AI Chat “Plugins/Extensions” can vary from access to repositories and tools like SalesForce, DropBox, and many, many more. That’s on the private sector side. On the government sector side, AI Chat plugins can range, some of which may require classified access, but all stem from a marketplace of extensibility for the AI Chatbots. That’s the real power of these chatbots. It’s not necessarily the knowledge of cheating on a university term paper. Educators are already adapting to OpenAPI, ChatGPT. A recent article in the MIT Technology Review, explains how teachers who think generative AI could actually make learning better.

Grassroots, Industry Standards Bodies should be driving the technology standards, and not lawmakers, at least until these standards bodies could expose all facets of AI Chat. Standards may also spawn from other areas of AI such as image/object recognition, and not all items brought about during the discovery phase should necessarily be restrictive. Some standards may positively grow the capabilities of AI solutions.

Chat Reactive versus Proactive Dialogs

We are still predominantly in a phase of reactive chat, answering our questions regarding the infinite. Proactive dialogs will help us by interjecting at the right moments and assist us in our time of need, whether we recognize it or not. I believe this is the scary bit for many folks who are engaging in this technology. Mix proactive dialog capabilities with Chat Plugins/Extensions with N capabilities/tools, creating a recipe for challenges that can be put beyond our control.

Riddle of the Sphinx: Improving Machine Learning

Data Correlations Require Perspective

As I was going to St. Ives,

I met a man with seven wives,

Each wife had seven sacks,

Each sack had seven cats,

Each cat had seven kits:

Kits, cats, sacks, and wives,

How many were there going to St. Ives?

One.

This short example may confound man and machine. How does a rules engine work, how does it make correlations to derive an answer to this and other riddles?  If AI, a rules engine is wrong trying to solve this riddle, how does it use machine learning to adjust, and tune its “model” to draw an alternate conclusion to this riddle?

Training rules engines using machine learning and complex riddles may require AI to define relationships not previously considered, analogously to how a boy or man consider solving riddles.  Man has more experiences than a boy, widening their model to increase the possible answer sets. But how to conclude the best answer?  Question sentence fragments may differ over a lifetime, hence the man may have more context as to the number of ways the question sentence fragment may be interpreted.

Adding Context: Historical and Pop Culture

There are some riddles thousands of years old.  They may have spawned from another culture in another time and survived and evolved to take on a whole new meaning.  Understanding the context of the riddle may be the clue to solving it.

Layers of historical culture provide context to the riddle, and the significance of a word or phrase in one period of history may wildly differ.  When you think of “periods of history”, you might think of the pinnacle of the Roman empire, or you may compare the 1960s, the 70s, 80s, etc.

Asking a question of an AI, rules engine, such as a chatbot may need contextual elements, such as geographic location, and “period in history”, additional dimensions to a data model.

Many chatbots have no need for additional context, a referential subtext, they simply are “Expert Systems in a box”.  Now digital assistants may face the need for additional dimensions of context, as a general knowledge digital agent spanning expertise without bounds.

 Sophocles: The Sphinx’s riddle

Written in the fifth century B.C., Oedipus the King is one of the most famous pieces of literature of all time, so it makes sense that it gave us one of the most famous riddles of all time.

What goes on four legs in the morning, on two legs at noon, and on three legs in the evening?

A human.

Humans crawl on hands and knees (“four legs”) as a baby, walk on two legs in mid-life (representing “noon”) and use a walking stick or can (“three legs”) in old age.

A modern interpretation of the riddle may not allow for the correlation and solving the riddle.  As such “three legs”, i.e. a cane, may be elusive, as we think of the elderly on four wheels on a wheelchair.

In all sincerity, this article is not about an AI rules engine “firing rules” using a time dimension, such as:

  • Not letting a person gain entry to a building after a certain period of time, or…
  • Providing a time dimension to “Parental Controls” on a Firewall / Router, the Internet is “cut off” after 11 PM.

Adding a date/time dimension to the question may produce an alternate question. The context of the time changes the “nature” of the question, and therefore the answer as well.

Amazon and Microsoft Drinking their own AI Chatbot Champagne?

A relatively new medium of support for businesses small to global conglomerates becomes available based on the exciting yet  embryonic [Chabot] / Digital Agent services.   Amazon and Microsoft, among others, are diving into this transforming space.  The coat of paint is still wet on Amazon Lex and Microsoft Cortana Skills.   MSFT Cortana Skills Kit is not yet available to any/all developers, but has been opened to a select set of partners, enabling them to expand Cortana’s core knowledge set.  Microsoft’s Bot Framework is in “Preview”  phase.  However, the possibilities are extensive, such as another tier of support for both of these companies, if they turn on their own knowledge repositories using their respective Digital Agents [Chabot]  platforms.

Approach from Inception to Deployment

  • The curation and creation of knowledge content may occur with the definition of ‘Goals/Intents’ and their correlated human utterances which trigger the Goal Question and Answer (Q&A) dialog format.  Classic Use Case.  The question may provide an answer with text, images, and video.
  • Taking Goals/Intents and Utterances to ‘the next level’ involves creating / implementing Process Workflows (PW).    A workflow may contain many possibilities for the user to reach their goal with a single utterance triggered.  Workflows look very similar to what you might see in a Visio diagram, with multiple logical paths. Instead of presenting users with the answer based upon the single human utterance, the question, the workflow navigates the users through a narrative to:
    • disambiguate the initial human utterance, and get a better understanding of the specific user goal/intention.  The user’s question to the Digital Agent may have a degree of ambiguity, and workflows enable the AI Digital Agent to determine the goal through an interactive dialog/inspection.   The larger the volume of knowledge, and the closer the goals/intentions, the implementation would require disambiguation.
    • interactive conversation / dialog with the AI Digital Agent, to walk through a process step by step, including text, images, and Video inline with the conversation.  The AI chat agent may pause the ‘directions’ waiting for the human counterpart to proceed.

Future  Opportunities:

  • Amazon to provide billing and implementation / technical support for AWS services through a customized version of their own AWS Lex service?   All the code used to provide this Digital Agent / Chabot maybe ‘open source’ for those looking to implement similar [enterprise] services.
  • Digital Agent may allow the user to share their screen, OCR the current section of code from an IDE, and perform a code review on the functions / methods.
  • Microsoft has an ‘Online Chat’ capability for MSDN.  Not sure how extensive the capability is, and if its a true 1:1 chat, which they claim is a 24/7 service. Microsoft has libraries of content from Microsoft Docs, MSDN, and TechNet.  If the MSFT Bot framework has the capability to ingest their own articles,  users may be able to trigger these goals/intents from utterances, similar to searching for knowledge base articles today.
  • Abstraction, Abstraction, Abstraction.  These AI Chatbot/Digital Agents must float toward Wizards to build and deploy, and attempt to stay away from coding.  Elevating this technology to be configurable by a business user.  Solutions have significant possibilities for small companies, and this technology needs to reach their hands.  It seems that Amazon Lex is well on their way to achieving the wizard driven creation / distribution, but have ways to go.  I’m not sure if the back end process execution, e.g. Amazon Lambda, will be abstracted any time soon.

Evaluating fobi.io Chatbot Powered By Google Forms: AI Digital Agent?

Interesting approach to an AI Chatbot implementation.  The business process owner creates one or more Google Forms containing questions and answers, and converts/deploys to a chatbot using fobi.io.  All the questions for [potential] customers/users are captured in a multitude of forms.  Without any code, and within minutes, an interactive chatbot can be produced and deployed for client use.

The trade off for rapid deployment and without coding is a rigid approach of triggering user desired “Goal/Intents”.  It seems a single goal/intent is mapped to a single Google Form.  As opposed to a digital agent, which leverages utterances to trigger the user’s intended goal/intent.  Before starting the chat, the user must select the appropriate Google Form, with the guidance of the content curator.

Another trade off is, it seems, no integration on the backend to execute a business process, essential to many chatbot workflows. For example, given an Invoice ID, the chatbot may search in a transactional database, then retrieve and display the full invoice.  Actually, I may be incorrect. On the Google Forms side, there is a Script Editor. Seems powerful and scary all at the same time.

Another trade off that seems to exist, more on the Google Forms side, is building not just a Form with a list of Questions, but a Consumer Process Workflow, that allows the business to provide an interactive dialog based on answers users provide.  For example, a Yes/No or multichoice answer may lead to alternate sets of questions [and actions].  It doesn’t appear there is any workflow tool provided to structure the Google Forms / fobi.io chatbot Q&A.

However, there are still many business cases for the product, especially for small to mid size organizations.

* Business Estimates – although there is no logic workflow to guide the Q&A sessions with [prospective] customers, the business still may derive the initial information they require to make an initial assessment.  It seems a Web form, and this fobi.io / Google Forms solution seems very comparable in capability, its just a change in the median in which the user interacts to collect the information.

One additional note, Google Forms is not a free product.  Looks like it’s a part of the G Suite. Free two week trial, then the basic plan is $5 per month, which comes with other products as well.  Click here for pricing details.

Although this “chatbot” tries to quickly provide a mechanism to turn a form to a chatbot, it seems it’s still just a form at the end of the day.  I’m interested to see more products from Zoi.ai soon

Evaluating Amazon Lex – AI Digital Agent / Assistant Implementation

Evaluating AI chatbot solutions for:

  • Simple to Configure – e.g. Wizard Walkthrough
  • Flexible, and Mature Platform e.g. Executing backend processes
  • Cost Effective and Competitive Solutions
  • Rapid Deployment to XYZ platforms

The idea is almost anyone can build and deploy a chat bot for your business, small to midsize organizations.

Amazon Lex

Going through the Amazon Lex build chat process, and configuration of the Digital Assistant was a breeze.  AWS employs a ‘wizard’ style interface to help the user build the Chatbot / Digital Agent.  The wizard guides you through defining Intents, Utterances, Slots, and Fulfillment.

  • Intents – A particular goal that the user wants to achieve (e.g. book an airline reservation)
  •  Utterances – Spoken or typed phrases that invoke your intent
  • Slots – Data the user must provide to fulfill the intent
  • Prompts – Questions that ask the user to input data
  • Fulfillment – The business logic required to fulfill the user’s intent (i.e. backend call to another system, e.g. SAP)

Amazon Lex Chabot
Amazon Lex Chabot

The Amazon Lex Chatbot editor is also extremely easy to use, and to update / republish any changes.

Amazon Chat Bot Editor
Amazon Chat Bot Editor

The challenge with Amazon Lex appears to be a very limiting ability for chatbot distribution / deployment.  Your Amazon Lex Chatbot is required to use one of three methods to deploy: Facebook, Slack, or Twilio SMS.  Facebook is limiting in a sense if you do not want to engage your customers on this platform.   Slack is a ‘closed’ framework, whereby the user of the chat bot must belong to a Slack team in order to communicate.  Finally, Twilio SMS implies use of your chat bot though a mobile phone SMS.

Amazon Chatbot Channels
Amazon Chatbot Channels

 

I’ve reached out to AWS Support regarding any other options for Amazon Lex chatbot deployment.  Just in case I missed something.

Amazon Chatbot Support
Amazon Chatbot Support

There is a “Test Bot” in the lower right corner of the Amazon Lex, Intents menu.  The author of the business process can, in real-time, make changes to the bot, and test them all on the same page.

Amazon Chatbot, Test Bot
Amazon Chatbot, Test Bot

 

Key Followups

  • Is there a way to leverage the “Test Bot” as a “no frills” Chatbot UI,  and embed it in an existing web page?  Question to AWS Support.
  • One concern is for large volumes of utterances / Intents and slots. An ideal suggestion would allow the user a bulk upload through an Excel spreadsheet, for example.
  • I’ve not been able to utilize the Amazon Lambda to trigger server side processing.
  • Note: there seem to be several ‘quirky’ bugs in the Amazon Lex UI, so it may take one or two tries to workaround the UI issue.

IBM Watson Conversation also contends for this Digital Agent / Assistant space, and have a very interesting offering including dialog / workflow definition.

Both Amazon Lex and IBM Watson Conversation are FREE to try, and in minutes, you could have your bots created and deployed. Please see sites for pricing details.

Microsoft to Release AI Digital Agent SDK Integration with Visio and Deploy to Bing Search

Build and deploy a business AI Digital Assistant with the ease of building visio diagrams, or ‘Business Process Workflows’.  In addition, advanced Visio workflows offer external integration, enabling the workflow to retrieve information from external data sources; e.g. SAP CRM; Salesforce.

As a business, Digital Agent subscriber,  Microsoft Bing  search results will contain the business’ AI Digital Assistant created using Visio.  The ‘Chat’ link will invoke the business’ custom Digital Agent.  The Agent has the ability to answer business questions, or lead the user through “complex”, workflows.  For example, the user may ask if a particular store has an item in stock, and then place the order from the search results, with a ‘small’ transaction fee to the business. The Digital Assistant may be hosted with MSFT / Bing or an external server.  Applying the Digital Assistant to search results pushes the transaction to the surface of the stack.

Bing Chat
Bing Digital Chat Agent

Leveraging their existing technologies, Microsoft will leap into the custom AI digital assistant business using Visio to design business process workflows, and Bing for promotion placement, and visibility.  Microsoft can charge the business for the Digital Agent implementation and/or usage licensing.

  • The SDK for Visio that empowers the business user to build business process workflows with ease may have a low to no cost monthly licensing as a part of MSFT’s cloud pricing model.
  • Microsoft may charge the business a “per chat interaction”  fee model, either per chat, or bundles with discounts based on volume.
  • In addition, any revenue generated from the AI Digital Assistant, may be subject to transactional fees by Microsoft.

Why not use Microsoft’s Cortana, or Google’s AI Assistant?  Using a ‘white label’ version of an AI Assistant enables the user to interact with an agent of the search listed business, and that agent has business specific knowledge.  The ‘white label’ AI digital agent is also empowered to perform any automation processes integrated into the user defined, business workflows. Examples include:

  • basic knowledge such as store hours of operation
  • more complex assistance, such as walking a [perspective] client through a process such as “How to Sweat Copper Pipes”.  Many “how to” articles and videos do exist on the Internet already through blogs or youtube.    The AI digital assistant “curator of knowledge”  may ‘recommended’ existing content, or provide their own content.
  • Proprietary information can be disclosed in a narrative using the AI digital agent, e.g.  My order number is 123456B.  What is the status of my order?
  • Actions, such as employee referrals, e.g. I spoke with Kate Smith in the store, and she was a huge help finding what I needed.  I would like to recommend her.  E.g.2. I would like to re-order my ‘favorite’ shampoo with my details on file.  Frequent patrons may reorder a ‘named’ shopping cart.

Escalation to a human agent is also a feature.  When the business process workflow dictates, the user may escalate to a human in ‘real-time’, e.g. to a person’s smartphone.

Note: As of yet, Microsoft representatives have made no comment relating to this article.