The AI with a two track mind

last updated:
a black cat wearing a pointy hat in front of a birthday cake

The AI with a two track mind

by:
last updated:

Speaking to a group of analysts recently, a product marketer for one of those vendors at the forefront of generative AI app assistance suggested that without the use of real business data to back it, the technology in enterprise terms is “largely a parlour trick”. We agree.

This week, the UK hosted its’ “AI Safety Summit” at Bletchley Park – the home of the UK’s World War 2 code-breaking unit – resulting in 28 nations signing an eponymous declaration which resolved to – among other things, “support an internationally inclusive network of scientific research on frontier AI safety […] to facilitate the provision of the best science available for policy making and the public good.”

As the BBC reported, “The summit focuses on so-called “frontier AI” – by which ministers mean highly advanced forms of the tech with as-yet unknown capabilities.”. My summary, free from the impartiality that the BBC is duty bound to report under, would have instead been “technology that doesn’t exist yet and in all likelihood won’t ever exist in the form being speculated upon by many of the chancers in attendance.” As I’ve mentioned before, I take a great issue with the use of fear as a tactic to explain AI to the wider public, especially when as in this case, it’s doing so free from the shackles of reality to allow some in the room to spiral off into creating pulp fiction of a vaguely scientific variety. 

Meanwhile, back in the land of things that very much do exist, we’re beginning to see some of the generative AI products that were splashy and announced in the spring, arrive at the point of general availability with customers. Microsoft’s 365 Copilot arrived for select bands of licensees on November 1st. While it’s not the first generative AI or even generative AI called Copilot from Microsoft out there, being associated with the 365 brand (née Office), it’s arguably the most significant and certainly the most widely visible.

As I mentioned last time out, as these products emerge from behind the dry ice, between the billowing flags, and onto the field, their cost also comes into view. Those select licensees are only those with 300 seats and above (to add Copilot for all at that opening volume, you’re talking an initial list price commitment of $100,000 per annum, before discounts, etc).

Getting the workforce settling in those new seats and being productive is the next challenge and one where I fear that some of the overselling of this first generation of generative AI assistance will – without proper forethought and planning – provide a disappointment. 

It’s not that the technology itself is without use and a high degree of charm. It was my cat Alex’s birthday this week, so I asked the Microsoft Bing Image Creator (powered by DALL-E 3) to create what it thought her birthday party might look like. The result is so adorable that I’ve used one of the creations to illustrate this post. But in the real world, turning these promising capabilities into demonstrable benefits for enterprises is far more than provisioning seats, especially when aligning them with all the data already resident within the organization.

Speaking to a group of analysts recently, a product marketer for one of those vendors at the forefront of generative AI app assistance suggested that without using real business data to back it, the technology in enterprise terms is “largely a parlour trick.” We agree. We also know that, as a result, this is where the hard yards will start for organizations that are committed to the use of the technology (and, as such, have written a report that seeks to outline the starting points to head in the right direction) and those yards will need to cover not only the technical and compliance wrangling over those data sets but also the involvement of the workforce whose roles could be significantly changed if the technology is successful.

If those governments in the room wished to contribute to something that has a demonstrable business impact in the respective countries – but at the same time insist it must be seen to be doing something AI related – providing the necessary knowledge infrastructure to deal with the AI of 2023 and how to deploy it with a positive organizational outcome for all, would be a far better use of their collective air miles. 

If they’d like to know where to start, we’d be more than happy to help.

Leave a Comment

We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.

Work Intelligence Market Analysis 2024-2029