SoylentNews

Arthur T Knackerbracket [soylentnews.org] writes:

--- --- --- --- Entire Story Below - Must Be Edited --- --- --- --- --- --- ---

Arthur T Knackerbracket has processed the following story [theregister.com]:

Designed to more efficiently perform the jobs of the human crew, the M5 [imdb.com] indeed did those jobs very well yet with such a terrifying lack of understanding it had to be disabled. But not before exacting a terrible price.

Last week, Microsoft 365 Copilot, a copilot, if you will, for the technology enterprise sold as performing human tasks with more efficiency, increased its prices by 5 percent [theregister.com], the first of many finely judged increments in the old style. Unlike the M5, it isn't in the business of physical destruction of the enemy, instead producing commercial victory with the photon torpedo of productivity and the phaser bolts of revitalized workflow.

Satya Nadella, Microsoft's own Captain Kirk, said at the end of October that AI was due to deliver $10 billion revenue [theregister.com] in the next quarter. It must be working, but in the absence of any clearly proven competitive results, it is time to look at Microsoft 365 Copilot itself 20 months after its introduction. Do the scanners show the utopian outcomes beloved of both Trek and Redmond franchises?

Some time back, this columnist noted the stark disparity [theregister.com] between the hype of the metaverse in business and the stark, soulless hyper-corporate experience. Line-of-business virtual reality has two saving graces over corporate AI. It can't just appear on the desktop overnight and poke its fingers into everything involved in the daily IT experience. Thus it can't generate millions in licensing at the tick of a box. VR is losing its backers huge amounts of money that can't be disguised or avoided, but corporate AI is far more insidious.

As is the dystopia it is creating. Look at the key features by which Microsoft 365 Copilot is being sold.

Pop up its sidebar in Loop or Teams, and it can auto-summarize what has been said. It can suggest questions, auto-populate meeting agendas. Or you can give it key points in a prompt and it will auto-generate documents, presentations, and other content. It can create clip art to spruce up those documents, PowerPoints, and content.

How is this sold? That it will make you look more intelligent by asking Copilot to suggest a really good question while doing an online presentation or a Teams meeting. What's also implied but unsaid: If you're the human at the end of this AI-smart question and want to look smart enough to answer it, who are you gonna call? Copilot.

The drive is always to abdicate the dull business of gathering data and thinking about it, and communicating the results. All can be fed as prompts to the machine, and the results presented as your own.

And so begins a science-fiction horror show of a feedback loop. Recipients of AI-generated key points will ask the AI to expand them into a document, which will itself be AI key-pointed and fed back into the human-cyborg machine a team has become. Auto-scheduled meetings will be auto-assigned, and will multiply like brainworms in the cerebellum. The number of reports, meetings, presentations, and emails will grow inexorably as they become less and less human. Is the machine working for us, or we for the machine?

Generative AI output feeding back into itself can go only one way, but Copilot in the enterprise is seemingly designed to amplify that very process. And you have to use it if you want to keep up with the perceived smartness and improved productivity of your fellow workers, and the AI-educated expectations of the corporate structure. 

Even the seemingly innocuous business of Copiloting the Dall-E-powered Designer graphics [microsoft.com] tool into documents is vile poison. The pictures generated can have no insights beyond what the generative prompts contain, diluted by whatever fantasies exist in the heart of the machine. But those pictures in a report or a presentation trigger expectations of extra meaning, of explication. The reason clip art feels so empty, so tiring, is that it speaks up our cognition and delivers nothing. That's human-designed clip art, taking some effort to find and include; the machine version is there instantly. It will push those micro-comas of PowerPoint viewing into new realms of unconscious revolt.

Human cognition is not the limiting factor of workflow. Under-stimulated boredom, tedium-induced laziness, performing actions that add nothing, and fear of failure-powered mediocrity are the enemies. All these find very comfortable new berths in corporate AI.

It is taboo to say how far your heart sinks when you have to create or consume the daily diet offered up in company email, Teams, meeting agendas, and regular reports. You won't be able to say how much further it will sink when all the noise is amplified and the signal suppressed by corporate AI. Fair warning: Buy the bathysphere [youtube.com] now.

There is an escape hatch. Refuse. Encourage refusal. When you see it going wrong, say so. A sunken heart is no platform for anything good personally, as a team, or as an organization. Listen to your humanity and use it. Oh, and seek out "The Ultimate Computer" – it's clichéd, kitsch, and cuts to the bone. The perfect antidote for vendor AI hype. ®

Original Submission