Enterprise hits and misses - are privacy regulations and distrust slowing innovation, and is gen AI set to disrupt IT services?

Jon Reed Profile picture for user jreed March 11, 2024
This week - data privacy comes into focus as the data breaches roll on (this time, it's United Healthcare). Will gen AI disrupt IT services? Or is the coding impact overrated? As always, your weekly whiffs.


Lead story - are privacy regulations and distrust slowing innovation?

This week, George examined two impediments to innovation. Let's start with distrust, via Is distrust slowing innovation? Why personal digital sovereignty could help.

Is there such a thing as an 'innovation distrust tax,' impeding legitimate societal improvements? George writes:

Distrust in the status quo may also discourage the wider sharing of information, slowing economic growth and impeding progress toward sustainability goals. This distrust tax on innovation occurs because individuals and businesses have little visibility or control over how their data may be used against their interests after sharing.

Smarter regulations may be part of the answer - but can individual data sovereignty help here?

In narrow use cases with simple controls, individual data sovereignty could make a lot of sense to #acceleratetrust by driving the useful data pool for training better AI for all of us. For example, I can imagine people contributing their medical data to help people like them, with the assurance that it could not be used to increase insurance rates.

Agreed - data transparency for individuals (and organizations) will certainly help with the opt-ins needed to make AI better, more ethical and more relevant. However, when it comes to AI in particular, I would point out that some mistrust of today's gen AI is warranted and healthy; there are specific problems with LLM output reliability that increasing parameter scale does not appear to solve.

Then we move onto the privacy question: do privacy regulations hinder innovation? George takes this on in What’s the privacy tax on innovation? No easy answers, but especially in the US, the state versus federal dynamic thickens the plot:

An emerging challenge for US companies will be a patchwork of different privacy frameworks that complicate compliance and the data and IT infrastructure controls required to support different rules. Anthony Cammarano, VP of Security, Privacy & Strategy at Protegrity, says it will incur a huge expense for businesses to do the same task in different ways.

George concludes with an 7 point plan for the next Uber or Airbnb, that wants to achieve a growth surge without violating privacy precepts. In terms of the legislative impact, the next step is to read George's US Executive Order promises privacy progress, which should further expose the pondscummy questionable role of data brokers in our lukewarm attempts at legally defining digital privacy.

diginomica picks - my top stories on diginomica this week

Vendor analysis, diginomica style. Here's my three top choices from our vendor coverage:

A few more vendor picks, without the quotables:

Jon's grab bag - A couple of notable International Women's Day posts:

One standout quote from Madeline's piece:

To achieve success as a female tech founder, Language I/O CEO Heather Shoemaker faced years of rejection and frustration. Her advice for other women following in her wake is simple - don't give up. Hopefully this message filters through, as the industry certainly can’t afford to lose more women.

Clearly, we've got plenty to work on before next year's IWD. On another front, Barb continued a diginomica theme on AI and creativity with How does AI support creativity? A marketing perspective. As readers know, this topic pushes so many of my buttons that I light up like a pinball machine. But I was surprised to read an entirely sensible viewpoint. Barb's interview subject James Manderson made no auto-magical claims about gen AI writing jargon-laden wonderful blog content:

Manderson offers several ways creativity and strategy work better together with AI, most of which have to do with its ability to analyze vast amounts of data:

'AI can process large volumes of data, providing valuable insights into customer behavior, preferences, and trends. These insights can guide creative decisions, ensuring they are both targeted and relevant. For instance, AI can determine which content types are most appealing to different customer segments, leading to the creation of more effective marketing campaigns.'

Hmm, I've made it this far without an AI chip being knocked off my shoulder - surely that won't hold...

Best of the enterprise web

Waiter suggesting a bottle of wine to a customer

My top six

  • United Healthcare's ransomware attack shows why supply chains are under siege - Louis Columbus looks at a concerning ransomware attack, with the fate of individuals' medical data still uncertain: "Peter Silva, Ericom, Cybersecurity Unit of Cradlepoint, told VentureBeat that 'the breach underscores how important it is to secure third-party vendors across a healthcare supply chain. Unfortunately, the real victims are the patients who are unable to get their medicine and now face the risks of having their health data compromised through identity theft.'"
  • Outside-in Process Q&A – Lora Cecere with another strongly-worded post on how to separate supply chain contenders from contenders.
  • India’s bloated IT services firms must learn from their startups to avoid GenAI meltdown - Phil Fersht of HfS Research issues a fresh warning for the legacy IT services industry: "For example, one enterprise has just concluded it can remove 15% of its IT staff from application testing by running an LLM against its testing processes. And this is just the start. When you consider that 20-30% of revenues for Indian-heritage IT services providers involve testing and quality assurance, it’s clear that there is growing pressure to weave GenAI into bread-and-butter IT areas like testing, routine maintenance and development." This piece, however, would be even stronger if it acknowledged that LLM coding output isn't perfect, nor is its ability to fully test and validate enterprise-grade code. Service and development efficiencies are, buzzword-alert, low-hanging fruit of gen AI, but the next phases of gen AI gains will be tougher. That said, the IT service industry has been in dire need to re-invention for years, it is overdue, and genAI is another big shove.
  • Can AI be a team player in collaborative software development? For views on AI development that navigate the pros and the cons, nobody has done a more consistent job than Joe McKendrick: "In the view of some industry leaders and experts, GenAI's impact doesn't seem to quite measure up to the miracles presented for business roles. 'Not every developer sees an opportunity in these tools for software development,' says Amrit Jassal, chief technology officer at Egnyte. 'There are quite a few rough edges, especially for experienced developers.'"
  • Lessons learned from four Zoho customers - Larry Dignan breaks out the high points from four Zoho customer interviews, via the Constellation Research team.
  • Two years later, deep learning is still faced with the same fundamental challenges - the difference from when Gary Marcus first posted his deep learning critique is how much has changed. Even the biggest gen AI evangelists are backing off from claiming scale will solve output accuracy problems. Top deep learning scientists are openly advocating/pursuing fresh approaches, inside and outside neural nets. Does this debate matter to enterprises? Yes - even though enterprise gen AI use cases are narrower and more controlled, the scientific obstacles impact the pace of innovation. Understanding this scientific debate will help us evaluate the occasionally overenthusiastic vendor proclamation - as well as why the consumer LLMs keep falling into the PR bogpit.

Overworked businessman


Honestly, the more I see of this hipster McDonald's uniform prototype, the more I like it:

Maybe I'm just missing my old skateboard... Keeping up with our data privacy theme:

I know Tesla says 'nothing to see here,' but it does appear problematic when access to a wifi log in means controlling your car:

I like my 'dumb' car more and more each year...

If you find an #ensw piece that qualifies for hits and misses - in a good or bad way - let me know in the comments as Clive (almost) always does. Most Enterprise hits and misses articles are selected from my curated @jonerpnewsfeed.

A grey colored placeholder image