A Room Where Executives Go to Get Help from IBM’s Watson
technologyreview.comSounds like a very cool demo, but the optimism of some of these ideas is quaint.
If Watson is good enough at speech and context recognition to accomplish these things IBM can/will sell truckloads of Watson as a Service for the purpose of monitoring employees.
WaaS will read everyone's email and parse conversations recorded by their desktops and phones to identify people who are off-task, leaking information or talking about unions.
That's the sort of thing Executives care about it. When I set up email retention systems they were primarily interested in being exempt from journaling. When I set up physical security systems they wanted to be sure the executive board room cameras weren't mic'd and faced away from the main presentation area.
In practice, both systems were primarily used to keep tabs on employees - who was dumb enough to send an email to the news from work and who's leaving early.
This is how 'intelligent' systems will be used - electronic overseers with distributed eyes and ears - long before it's confined as a guest boardroom showpiece that gets tossed for suggesting the CXO get off his soapbox in a timely manner or correcting his knowingly incorrect assertions.
And yet we have millions (literally) of companies who do not employ this type of monitoring, despite it being trivial from within google apps.
> WaaS will read everyone's email and parse conversations recorded by their desktops and phones to identify people who are off-task, leaking information or talking about unions.
This specifically is not a terribly difficult monitoring task. You don't need something like watson to do it. Again, already available, still not heavily deployed.
The truth is people mostly trust each other and don't do shit like that. At some places they will, but watson is not the enabler here.
>'The truth is people mostly trust each other and don't do shit like that. At some places they will, but watson is not the enabler here.'
I'm not sure we're talking about the same things here?
Perhaps in the startup world that is common on HN things are small, lean and trusting enough that this sort of outlook is possible.
The context of the article, Executives at the sort of companies that buy IBM, they absolutely invest in monitoring and retention by choice or by law.
Where I'm sitting now we retain every email, voicemail and internal chat for tens of thousands of people and video from hundreds of cameras. We retain 100% of voice calls for some sections. People who deal with the public have all their calls recorded as well as their screens. Every meaningful door is operated with an access card.
The limiting factor on implementation of these things is not some non-specific altruism, it's money. Money for licensing, infrastructure and operations - it's not cheap, not close.
No, Watson is not the specific enabler for these things to happen as they clearly already do, but the sort of intelligent XaaS that Watson could be is exactly what would make these systems cheaper and more effective.
Responsiveness for specific categories could be determined on the fly leading to lower size of retention and operators/administrators could probably be eliminated entirely.
My point is that the analysis of this data is not particularly cost prohibitive. If you want to take all the emails you have recorded and figure out if people are talking about unions you really really do not need something like Watson to do it. I'm not saying people are or are not recording this data, I'm saying they don't have an interest in really digging through it, and if they did they would without something like Watson.
TLDR: Watson has nothing to do with monitoring + retention.
>'My point is that the analysis of this data is not particularly cost prohibitive. If you want to take all the emails you have recorded and figure out if people are talking about unions you really really do not need something like Watson to do it.'
Clearly, but for whatever reason you've decided that 'email' alone is relevant here, not the bits 'speech and context recognition' and 'parse conversations recorded by their desktops and phones' which are the obvious applications of something like Watson.
>'I'm saying they don't have an interest in really digging through it.'
My direct experience says you are simply uninformed about this.
>'if they did they would without something like Watson.'
Only in the context of the 'emails' strawman you've constructed above.
TLDR: You are missing or deliberately avoiding the point. Voice recording and analysis is already big business, but speech and speaker recognition are still lacking and context awareness is obviously dependent on both. That is the exactly the sort of thing that this article specifically demonstrates Watson as being excellent at.
Why would people who are actively digging through phone calls not be digging through the relative low hanging fruit of emails? I'm thus using lack of interest in digging through emails as a proxy for lack of interest in this kind of analysis - is that unreasonable?
Edit: And to be clear: I'm not saying there isn't a market for this product, just that the concept that it would become the dystopian norm is a bit melodramatic.
Just having software that can record and transcribe only relevant conversations, record action items, and email summaries would vastly improve the state of the meetings I attend.
Nothing worse than wasting 3 hours in a meeting and coming out where participants remember things incorrectly and go off in different directions.
Indeed but so far, software has shown itself much more capable of solving logical puzzle than determining relevance. It seems fairly clear that the Jeopardy win came from the program having a lot of data and logical statements from which it could make relevant guesses and deductions.
If the software could make a coherent summary of and answer questions about what took place in a single, informal meeting, that would be an incredible advance. I haven't seen evidence of this.
Right, at the company I currently intern at this is all done manually by the software librarian. It seems like there is a lot of room for automation in meetings; bonus points if such a system could tie into the issue tracker/SCM (whatever that may be, we use Accurev). Although Watson sounds like it would be helpful in cutting down on the number of pointless arguments in meetings where not everyone is on the same page.
Sounds like you have some poor chair's of meetings. Minutes should be done by a secretary/note taker or a PA.
And if you think that's old school you have never had the luxury of meeting where you had a professional PA it makes an order of magnitude difference.
>“I recommend eliminating Kawasaki Robotics.” When Watson was asked to explain, it simply added. “It is inferior to Cognilytics in every way.”
Getting some serious M5 vibes right now.
What's M5?
I had a short gig at Cognilytics in 2010, and I can't figure out what any version of it has to do with Kawasaki Robotics, but...
That's a nice recommendation, but it's good to have backup choices even if they have no relative advantage economically.
IBM reps have been pitching Watson to me for years, but never have we seen an implementation or a commercial offer, much less a demonstration of something of value to us. Weirdly it clearly exists, or did exist, but I think marketing and sales have robbed R&D of the budget required to create a product and until they sell the non existent I think this will continue to be the case. If I was an IBM stockholder I would be pissed off.
>IBM reps have been pitching Watson to me for years, but never have we seen an implementation or a commercial offer, much less a demonstration of something of value to us.
What exactly have they been pitching you?
Watson for CRM.
The pitches are at the level of "a pipeline off tools" + "a new level of automation", getting to detail elicits a "call with the team" which involves a discussion about "the pipeline of tools" and a possible POC, which has not happened.
We actually had a Question Answering system research strand ourselves which we couldn't really get anywhere with and I killed (mea-culpa) because of Watson, Google and Siri. I thought that we had no chance of making any headway against such strong competition and imagined that QA tech would be available in the contact centre and on the Intranet in short order.
More fool me! I've been embarrassed ever since.
Sounds like they don't really know what Watson would be best for and are hoping their customers can tell them. Not a terrible strategy but really doesn't look right coming from IBM.
> IBM’s researchers are also considering other ways the technology at work in their current demo might help out in a workplace—for example, by having software log the relative contributions of different people to a discussion
I find that incredibly creepy and hope to god that never is developed.
If it records the -actual- contribution and not the number of words spoken I'd be all for it. I'd also want it to measure the amount of hot air people talk.
That's what I was thinking about. If Watson's cognition is as good as the article suggests, it can easily detect when one silent fellow speaks up once during the whole meeting and changes the overall flow of conversation completely.
If technology like this really had a basis in reality, one would think that some early prototype would be delivering positive business results.
When I think IBM, I think accounting tricks to maximize stock price. Not seeing deep insight.
I wonder if they pick up a phone and ask "Come here Watson I need you."
"Multivac, I have a question..."
The failure (or limitations, because I don't wish to talk down the very real achievements) of current AI is the overfocus on building computers that provide answers to our questions.
We're not going to get real AI until we develop a system that asks us questions and has a sense of curiosity. A system that can make suggestions is excellent, but as described it's effectively cybermancy (http://en.wikipedia.org/wiki/Methods_of_divination). One doesn't get the sense that Watson is ever going to interrupt or pose a question on its own initiative, other than to clarify a human request put to it.
In my mind, this is a feature and not a bug. The greatest cybermancy spells technologists can create are inherently safer than an AI with potential to go rogue. Why do we want to create a god when we can become gods ourselves?
We're not going to become gods by making simulacra of perfectly obedient servants. If anything, this is more likely to support the development of a hostile AI that considers total submission a desirable end state to impose on others.
This has nothing to do with AGI, period. Watson is certainly not going to rebel against you. Nor is it intended to be significant progress towards human level AI.
I also find it funny you think the secret sauce to AGI is "a sense of curiosity". Yes, as if that was the only thing holding computers back. If only someone had thought of that before.
Of course, I put all my thinking about AGI into those two sentences and now you have crushed my dreams forever.
"Starving minds, welcome to Dr. Know! Where fast food for thought is served up 24 hours a day, in 40,000 locations nationwide. Ask. Dr. Know - there's nothing I don't."
Reminds me of MUTHER from the Venture Brothers...
Sure, but is Watson a Six Sigma Green Belt?
Coming from IBM, I am sure it is green. :-)