RE: re Findable Help proposed success criteria

I am afraid I have come to this topic a month later but I have been looking into the issues around chatbots and the problem with their replies to user's input.

What happens when they are not providing the expected type of answer or saying they do not understand the question.  What impact does this have on the user?   A sense of failure, confusion or frustration etc.

I am wondering if you received any support Jennie for your query as I am worried that perhaps Chatbots can offer wonderful support in certain circumstances if the user realises their limitations.  However, they can be disastrous if the user is unaware of the fact that a human may not be answering their questions.  This link is not an academic one but makes some good suggestions https://freshdesk.com/customer-support/chatbots-vs-humans-blog/


IBM have an entire set of masterclasses on the subject https://tinyurl.com/t4wk5j9


Best wishes
E.A.

Mrs E.A. Draffan
WAIS, ECS , University of Southampton
Mobile +44 (0)7976 289103
http://access.ecs.soton.ac.uk<https://www.outlook.soton.ac.uk/owa/redir.aspx?C=69b1RzNTDwem3wbm4pLRmuYfTLt16YjcghtEpZBsF5Sebx78I2DUCA..&URL=http%3a%2f%2faccess.ecs.soton.ac.uk%2f>

From: John Foliot [mailto:john.foliot@deque.com]
Sent: 30 October 2019 15:25
To: Delisi, Jennie (MNIT) <jennie.delisi@state.mn.us>
Cc: public-cognitive-a11y-tf <public-cognitive-a11y-tf@w3.org>
Subject: Re: re Findable Help proposed success criteria

Hi Jennie,

The concern here (at least for me) is defining "effective", which is and always will be subjective, right down to the individual end-user.

There is also the additional concern of why the chat-bot is there, and what task or activity might be associated to that particular form of "help" - i.e. the complexity of the task (for any individual) will have a direct relationship on needing help and/or assistance, which might range from an automated "FAQ" (i.e. one potential chatbot solution), versus at the end, a user *absolutely* needing to have real-time human assistance.

In other words, there are different types (and levels) of "help", and if the actual goal here is to provide real-time human assistance, then we need to be crystal clear in scoping that out, which is but one form of help.

JF

On Wed, Oct 30, 2019 at 9:57 AM Delisi, Jennie (MNIT) <jennie.delisi@state.mn.us<mailto:jennie.delisi@state.mn.us>> wrote:
Hello,
Yesterday I had the opportunity to discuss the Findable Help proposed success criteria with the Accessibility Guidelines group. I think it went well, and there was both good feedback and a greater understanding of the challenges faced by some with cognitive disabilities as they try to complete a task and need help.
I’m currently reviewing all the feedback, and will be reaching out with requests as I revise and prepare for Tuesday’s meeting when this will again be on the agenda.
One suggestion from a group member was “I think we need data to back up automated chatbots being ineffective, especially since they might also be used in support pages or help content.”
Another posed the concept of the best, most intuitive and helpful chatbot you could conceive of – what if someone makes that? Would it still fail this?
So, I’m asking this COGA group:

  1.  Please send links to research you may have in regards to chatbots not being effective.
  2.  Could you send a sentence or two about: If there was an ideal chatbot, it would…?
Please know that I am not confident that we could write enough specifics to make the chatbot piece doable for individuals with cognitive disabilities, however, I want to pose the question to this group since it was asked during the meeting.
Thanks,
Jennie
Jennie Delisi, MA, CPWA
Accessibility Analyst | Office of Accessibility
Minnesota IT Services | Partners in Performance
658 Cedar Street
St. Paul, MN 55155
O: 651-201-1135
Information Technology for Minnesota Government | mn.gov/mnit<https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmn.gov%2Fmnit&data=01%7C01%7Cead%40ecs.soton.ac.uk%7Caf0fb8c1219e4ed3570b08d75d4d7f76%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=oJEXCF6o7HIX0aoH51235INYTQ9pOWB0WSovypgFGZE%3D&reserved=0>
[Minnesota IT Services Logo]
[Facebook logo]<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.facebook.com%2FMN.ITServices&data=01%7C01%7Cead%40ecs.soton.ac.uk%7Caf0fb8c1219e4ed3570b08d75d4d7f76%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=Lv%2FHZTC6%2FtUC8JPEBtRqyd4p0fRrCYM9aTtOnP7UzHs%3D&reserved=0>[LinkedIn logo]<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.linkedin.com%2Fcompany%2Fmn-it-services&data=01%7C01%7Cead%40ecs.soton.ac.uk%7Caf0fb8c1219e4ed3570b08d75d4d7f76%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=RH0JXzpT%2Bky69pyBox6Na6fxXOHNpzyR7mrAps35PR0%3D&reserved=0>[Twitter logo]<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Ftwitter.com%2Fmnit_services&data=01%7C01%7Cead%40ecs.soton.ac.uk%7Caf0fb8c1219e4ed3570b08d75d4d7f76%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=o4FdkuFfujUBQrGNEn0u%2BJKR7izKxrFbf713F%2BrFMPI%3D&reserved=0>


--
​John Foliot | Principal Accessibility Strategist | W3C AC Representative
Deque Systems - Accessibility for Good
deque.com<https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fdeque.com%2F&data=01%7C01%7Cead%40ecs.soton.ac.uk%7Caf0fb8c1219e4ed3570b08d75d4d7f76%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=JLSGxsR4d8fWerKtGx3LgWK2p1aGPzWjYTjNAak29EI%3D&reserved=0>