Lead story – AI and sustainability: unresolved questions impact data center expansion
Can AI make a valid contribution to sustainability, and perhaps, as some have brazenly asserted, solve it own AI/sustainability dilemmas? Perhaps, but today is not that day. Cath takes up the issue in No silver bullet for AI and data center sustainability challenge – an integrated approach is needed. First, the problem statement:
To effectively address the problem, it will be necessary to tackle all facets of the sustainability challenge. These include e-waste, heat reuse, embodied carbon and product circularity. Today though, the focus is predominantly on cutting energy usage and moving to renewable sources.
Though energy usage is cited by one expert to jump 20x due to AI, there are caveats worth pursuing. One that Cath points out: a lack of creative thinking on how to address this. She quotes Devorah West, Senior Advisor for Climate and Energy at the Tony Blair Institute for Global Change:
It requires integrated thinking in what are currently siloed departments, which means we’re not there yet in the UK. Plans aren’t linked up. For example, with heat reuse, you have to think about things like where data centers are placed and where workloads could go. To do it effectively, government needs more creative thinking and to join the dots.
Cath expands on the challenges in Action is lagging behind good intentions in the transition to net zero:
So far, a respectable 68% of companies in the IT and communications space have come up with a plan to cut carbon emissions (compared with an average 57% among their peers).
However, a mere 35% have finished working out a full baseline carbon footprint covering Scope 1, 2 and 3 emissions (compared with a third of organizations across other sectors). The figure rises to 44% if those that have put together at least a partial baseline are included (compared with 52% elsewhere).
I lean skeptical that technology can solve the problems that technology helped to create. But that doesn’t mean we shouldn’t use every tool at our disposal, from tech to purposeful entrepreneurship to savvy policy making (no small feat), to close the gap. Cath concludes:
Concrete action appears to be lagging somewhat behind the good intentions. A lack of clear government policies combined with inadequate levels of company action mean that huge amounts of work remain to be done here. As Zie says, the tech industry has a valuable role to play in helping organizations move forward, but it needs to do more. And quickly.
Sounds right to me…
Vendor analysis, diginomica style. Earnings reports of note:
Here’s my three top choices from our vendor coverage:
- Team ’24 Europe – Atlassian channels AI in the context of enterprise teamwork – Phil issues his Atlassian event wrap: “There’s no doubt that, as Atlassian’s CMO, Zeynep Inanoglu Ozdemir, told us at Team ’24 Europe, ‘AI is actually changing the fabric of teamwork.’ But that doesn’t mean we have to throw out everything we’ve learned about enterprise teamwork and start over.”
- Never mind the copilots; make marketing the air-traffic controller! Consequential AI takes flight – Barb reviews Bloomreach’s approach to AI and marketing – I like this approach quite a bit more than most I’ve seen.
- Celonis launches AgentC – ‘Process Intelligence essential to any enterprise AI strategy’ – Derek updates on Celonis’ agentic AI plans, live from Munich: “As noted above, today Celonis has launched a suite of AI agent tools, integrations and partnerships, dubbed AgentC, that utilize the Process Intelligence Graph to provide enterprises with context-driven AI. First platform integrations include Microsoft Copilot Studio, IBM watsonx Orchestrate, Amazon Bedrock Agents, and open source developer environments like CrewAI.” Also see Derek’s use case: E.ON uses process intelligence provided by Celonis to support transition to cleaner energy.
UiPath Forward 2024 – diginomica coverage – does agentic AI make RPA irrelevant? No – and there are good reasons why. It’s all about the relative strengths of both, and why agents need ultra-reliable, RPA-type automations for enterprise subroutines that they can orchestrate, invoke, or invoke on a user’s behalf. Alyx and I were on the ground for UiPath’s next act, agentic AI. But how did that go over to customers heavily invested in UiPath’s RPA? Surprisingly well – due to the self-reflection and brain food offered up on stage, in lieu of what I call futuristic productgasms. Here’s our coverage to date, with more to follow:
A few more vendor picks, without the quotables:
The latest diginomica use cases:
Jon’s grab bag – George examines IBM’s hallucination mitigations in Open sourcing AI guardrails – IBM’s push to improve safety and reduce hallucinations – though in my view, we won’t get truly beyond the LLM accuracy issue until we move from context to meaning. But: semantical context, where machines have a far better grasp of how words and concepts relate, is a worthwhile endeavor.
Meanwhile, Chris takes another angle on one of his signature topics: the impact of AI on creativity. This is spot-on:
One real-world challenge is subtler and more insidious, perhaps, than the existential threats suggested by science fiction’s doom-mongers. And that is the very real danger that AI poses to human creativity – in particular to skilled humans’ ability to make money from their talents and expertise. (AI and creativity – don’t be seduced by dishonest AIs, urges Professor).
I don’t believe AI truly threatens human creativity in the brilliant/authentic sense of that word (some artists may even put AI through its paces for surprising effects), but AI absolutely threatens some creative professions. How we handle that will not only define modern creativity but also, perhaps, our own humanity. Also see: Chris’ AI and creativity – should music be the food of AI?
Best of the enterprise web
My top seven
Agentic AI without process optimization, orchestration will flop – Constellation’s Larry Dignan has another angle on the themes we picked up on at UiPath, re: why (probabilistic) gen AI agents need process execution tech in support. Otherwise you are just “orchestrating” chaos, or silos. Dignan’s take:
In recent weeks, the flow of agentic AI news hit a fever pitch and it appears that vendors have gone from launching AI agents in their platforms to catapulting to “agent of agents.” Many of these plans are short on process automation and orchestration. At this moment, agentic AI is a game of executing tasks autonomously within a vendor’s platform. First, you got the data silos. Then you got the dime-a-dozen copilots within your applications. And now you’re getting AI agents that aren’t going to operate across platforms and processes.
What to do about it? Dignan issues a list of actions, including a “neutral” vendor and process mining/orchestration. He writes:
Building AI agents won’t be a problem. Managing them will be.
There’s where we may diverge. Building agents won’t be a problem – heck, build twenty of them if you want this afternoon – but as, I’ve argued, building effective agents in an anterprise context won’t be easy. Accuracy, use case selection, figuring out exactly when to invoke humans, contending with the costs of outliers (sometimes they will be fine, sometimes they won’t), that’s all part of the AI agent game. Oh, I forgot, if you use the word “revolutionary” enough times, then no one will ask you about your project’s ROI.
- AI Won’t Fix Your Software Delivery Problems – a good one from The New Stack, as we get more useful clarity on how gen AI aids developers, and what its pros/cons are: “AI reduced throughput by 1.5% and stability by 7.2%. If perception of code quality is increasing with AI assistance but stability is dropping, it indicates a mismatch between how well we think it performs a task and how well it actually does.”
- Annual Planning Sucks — A CPO, CRO, CFO, and COO Share Advice on How to Make it Better – A meaty and timely piece from First Round.
- As The Winds Blow: Improving Supply Chain Planning – Lora Cecere doin’ work: “In 8 out of 10 companies, planning is an entry-level job. The turnover is high, and the quality of the plan suffers. Organizations with planning tenure outperform.”
- There Is Just One Way To Do Open Source Security: Together – Hacker One’s CEO has some strong words: “We need to practice sharing the bad news because that’s the only way to make good news.” If that means publicly reporting embarrassing security failures, so be it.”
- The Global Surveillance Free-for-All in Mobile Ad Data – Brian Krebs on the slippery slope of surveillance, mobile ads and rogue personalization.
- Debunking Hype: China Hasn’t Broken Military Encryption With Quantum – Quantum Computing will bring its share of new security risks, but inaccurate, viral reporting is also a risk – both addressed in this piece.
Whiffs
I’m not sure if health care is gen AI’s lowest-hanging-fruit. You?
Researchers say an AI-powered transcription tool used in hospitals invents things no one ever said https://t.co/5w2Is8kCBq
“The problems persist even in well-recorded, short audio samples.”
-> no problem, let’s just use it anyway, it’s just health care we can live with…
— Jon Reed (@jonerp) October 28, 2024
But repairing your own stuff is always a win, especially if it’s a broken McFlurry machine:
It Is Now Legal to Hack McFlurry Machines (and Medical Devices) to Fix Them https://t.co/9GmwGYTnst
-> not sure this McFlurry news is going to affect Western Civilization, but this is also good news for hospitals too…
— Jon Reed (@jonerp) October 28, 2024
Finally, I’m always here to solve everyone’s LinkedIn etiquette questions:
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.