Summary of SIM scenario-building exercise on the future of AI

3-to-4 dozen Society for Information Management CIOs weigh in

Nov 26, 2024

To set the table, this scenario-building exercise engaged 40-plus CIOs from a variety of industries. It was conducted over a single hour on day 3 of the “Tech Exec 24” conference at the Fairmont Princess Resort in Scottsdale AZ (20th November).

I designed the drill, leveraging Perplexity.ai throughout, so the scenarios presented basically represented AI predicting the future of AI, with my guidance.

I asked Perplexity to give me the two framing questions:

AI’s spread throughout society and economy
- Restricted (by public acceptance, government regulations, weak advances, etc.)
- Ubiquitous
The surrounding information security environment
- Vulnerable
- Robust.

I then asked for pun-centric movie titles for each of the four resulting scenarios. In each instance, I got a bunch of decent attempts, with me selecting the bit that really triggered something in my head and correcting/improving on that basis — to wit, The Silence of the Algorithms became The Silence of the LANs (both funnier and more descriptive of a balkanized IT world).

As usual, as I have found, Perplexity outperformed me on speed and gross content, but I beat Perplexity on refined, finished content.

The Centaur Solution prevails!

Once I settled on the four names, I turned to Chat GPT to generate the corresponding movie posters (my thanks to colleague Tom Zorc for that painstaking effort).

I then asked Perplexity to give me the 9 best descriptors of each of the four scenarios. As per my experience, about one-third of them I could use without alteration, another third needed alteration/improvement, and the last third were easily jettisoned and replaced by me. Whenever I generated one on my own, I ran in back through Perplexity to make sure it “resonated,” yielding clear responses that validated my choices.

That’s how I leveraged AI to build the scenarios and they struck me as very much what I would have expected of myself if I had put in several weeks of research in THE BEFORE TIME. So, honestly, it was like having this great research assistant who worked at blinding speed and — for the most part — got what I was going for.

So, here’s a summation, with first-cut analysis by me, of the scenario-building drill that we ran through Mentimeter.com (which performed nicely).

WARM-UP: Describe AI in one word (unlimited entries)

As participants filed into the conference room, they were encouraged to go to Mentimeter.com, plug in the event code, and start off by offering us one-word descriptions of AI.

We got this word cloud in reply:

Clumping actual repeats, I came up with the following top performers:

EXCITING at 14 mentions
IMPACTFUL at 7.

Asking Perplexity to group, by theme, the total haul (with some trimming by me), I got the following breakdown, adding in the multiple-entries as I saw fit:

Positive Attributes (25 of 78, or 32%)

Amazing, awesome, empowering, exciting (x14), exhilarating, fascinating, fun, magic, mind-blowing, progressive, promising, and value-creator

Scope (18, or 23%)

Big, epochal, life-changing, game-changing (x2), impactful (x7), limitless (x2), transformative (x3), and ubiquitous

Challenges and Concerns (15, or 19%)

Cautionary, challenging, concerning, confusing, conundrum, dangerous, daunting, fallible, frightening, intimidating, overhyped, scary, stress, tired, and unknown

Growth and Evolution (12, or 15%)

Accelerator, automation, disruptive, evolution, expansive, explosive, exponential, inevitable, infancy, renaissance, revolutionary, and unbridled

Practical Applications (8, or 10%)

Data, efficiency, enabler, interface, learning, predictive, productivity, and strategy.

Add it all up and, by my reading, more optimism (positive attributes) than pessimism (challenges and concerns), but definitely an awed sense of scale (scope + growth and evolution). You get that feeling that we’re all standing behind the earth berm at the Trinity site and we’re going, Whoa!

Letting the group know what they were in for

The agenda as presented.

Distribution drill: spend a budget of 100 units

The idea here, based on outputs by Perplexity.ai, was to force these CIOs to rank-order by weight how they view the following AI industry challenges, meaning, if they had only $1 to spend against the eight issues, how would they spread that money around?

This is how they voted:

I was a little surprised here that Measuring Business Value was … I dunno … that much of a mystery to be solved. But, I guess it makes sense when you’re talking something pervasively transformative. Nobody really knows what’s actually possible, but, to avoid the hype/disillusionment cycle, the benefits need to be expressed in understandable business terms.

You can’t improve what you can’t measure, am I right?

“Integration” was THE buzz word of the conference: It’s here and it ain’t going away, so now’s the time to seriously integrate AI into what we already have.

Very rubber meets the road.

Next is governance concerns just topping data requirements, followed by a pair of essentially cost considerations (Implementation + Energy).

Then the two scoping elements bring up the rear: How to scale and how to gain the public’s trust — two highly interrelated dynamics, no?

But you get a sense of first-things-first here:

Figure out the business case
Leverage what we already have (integration)
Figure out the whole data-ownership thing (Exactly whose IP can we plunder without triggering too much government backlash?)
Then deal with the sheer limitations (costs)
Then run this up the public flagpole and see how high we can get it (Scale v. Public).

So, all in all, pretty sensible trajectory.

Now let’s play the role of AI’s Nostradamus …

This one was a true brain stretcher: all fairly explainable concepts ginned up by Perplexity.ai, except — and ironically enough — for the case of “Explainable AI, or XAI.”

But that’s actually pretty explainable (by AI):

Explainable AI (XAI) is a collection of tools and techniques that help people understand how and why machine learning models make decisions. It's also known as interpretable AI or explainable machine learning.

It’s basically the old mother/father-in-law test: If you really understand how something works, you ought to be able to explain it to a friendly (one hopes) but-non-expert acquaintance.

This was the sequencing as voted by our CIOs:

First place made sense to me, as did last place (truly out there, per an expert speaker that morning). I was a bit surprised to think that Predictive Threat Analysis Systems in Cybersecurity would come that fast, but, as we discussed, it kind of makes sense: Just like we find that the best way to defeat drones is with other drones, the best way to protect AI is to rely on it to predict and respond to threats to its safety.

I know, I know. Just saying that is a bit scary — the trust factor.

I was happy to see AI-Generated Personalized Education coming on that fast, for I see that as a huge fix to younger generations’ legit fears of being unprepared, education-wise, for this economic future.

Healthcare Therapies up next, hopefully coming before we all trust our lives and limbs to driverless cars (Elon, of course, would have set the latter mark for several years ago).

A bit unnerving to think that we have to live with these Black Box/trust me explanations (pre-XAI) for that long, but hey, at least it comes BEFORE Emotionally Intelligent AI.

I mean, it’s not funny if you have to explain it, right?

After that it’s Iron Man suits for all as we slip into Quantum AI heaven.

All in all, pretty sensible sequencing, by my limited understanding.

The next step?

Run that outcome back through AI to see if AI agrees!

Seriously … I’m going to do that.

Presenting the four Future of AI scenarios

Pretty basic stuff previewed above and in the slides already shared on this Substack.

First, start with two questions:

That gets us four outcomes, generically labeled.

With prompts from Perplexity and Chat GPT, I named them as follows:

I already told you my mod of “The Silence” scenario (LANs swapped in for Algorithms).

The lower right case started with Perplexity offering Phantom + various appropriate terms, but I wanted the first letter of the second term to match Menace (M), so I subbed in Malware, which works just fine.

On the upper left, I saw the word Sentience and the rest was easy for this Emma Thompson fan (age-appropriate movie star closest to my spouse in looks).

Ditto for the word Guardians for the upper right category.

Now, as I explained to the participants, I went with this traditional X-Y array of:

Worst case in the lower left/starting point
Best case in the upper right/downstream point
And then the two mixed cases offer differing degrees of speed + danger for the journey from worst-to-best cases.

Here, The Silence of the LANs is clearly the darkest in terms of progress and trust, so Local Area Networks is a great image: a siloed-off IT universe.

Opposing that in optimism is Guardians of the Gigabyte (I couldn’t come up with a good U word so went with alliteration — a poor man’s pun). Maybe I could have gone with Guardians of the UNIX-verse?

Geez. Now that I say that, I think I missed a golden opportunity there, one I am likely to fix in future iterations.

As for the two pathways from worst to best: The slow and careful route is neatly evoked by Sense and Sentience, giving off that Victorian, we’ve-got-it-under-control vibe, albeit with a societal calmness purchased at the price of equality and reach. But I liked that historical reference (even if the movie/book evokes the early 1800s and we’re really talking about the late 1800s, but, let’s be honest, most people aren’t going to catch that distinction).

The point is, the Victorian age was one of massive technological advance across a host of industries, and yet, at least in the case of Great Britain, social structures remained firm in their guard-rails-like role of processing all that transformative change. Did that come at some revolutionary cost down the road? Yup, but you get the idea.

As for the too fast + too scary pathway where technologies race ahead of security and the economics races ahead of the politics, The Phantom Malware, once the right Star Wars imagery was achieved in the movie poster (not easy, because the power of the Mouse is vast and nefarious), seems like a great fit. Like the Star Wars movie, there is this threat right out there in the open that nobody quite sees, much less recognizes: We are all being manipulated!

Next up I provided the following descriptors for each scenario pathway:

And then we were off to the races.

Headlines from the Future: The Silence of the LANs

I threw up the following reminder and the inputting began.

Here is the resulting dump:

Massive Recession Fueled by Lack of Trust in Institutions (could use some more causality, but I get the idea)
Distrust in AI Proven Right in Major Utilities Outage (almost a certainty, right?)
End of the World Near (right to the punchline)
Get Out! (cheating there by using another movie title, but I get the notion and it’s pretty good, actually)
What We Lost Today? (mournful essay, one imagines, in the New Yorker)
HALible Lecter is Here (trying too hard to be funny, are we?)
Public Loses Faith in Leaders and Institutions, Puts Trust in AI and Robots (dark enough for you?)
Everything is Fake, Trust Nothing (I already have that T-shirt)
Reversal of Fortune (hmm, intriguing but could have used a detailed subtitle)
Shh! (Jordan Peele’s next movie)
Massive Cyber Attack Takes Down Power Grid: More Dead Than 9/11 and Trillions Lost (Defense AI fails, or is it that Threat AI succeeds?)
Government Gives Up, Privatizes Cyber for Citizens! (the concept that AI is too important to be left to the government, I imagine)
Social Security Database Hacked; All US Citizen Data Now in ‘Wild’ (bit generic)
Network Scanner Determines Business Fraud (seems better off in the Best Case scenario)
Game Over! (somebody remembers Aliens, and I approve)
Cybercrime Gangs Take Ransomware to New Level, Crippling Business Innovation (pretty good)
Nuclear Power Plant Gets Hacked; Pandemonium Rampant (again with the fears of public utilities being vulnerable! I would like to dig into that notion as it relates to AI)
Russia Hacks and Controls US infrastructure, Holding Country Hostage (also pretty good)
Elon becomes Robot! (I’ll bite)
How AI Pushed Existence into the Matrix (this intrigues)
Top 1% Are Only People Able to Leverage AI (appropriate)
AI Completes First Robotic Surgery from Diagnosis Through Post-Operative Care (I would also put that one in the Best Case)
Trump Puts Brain in Robot, Earns Third Term (nobody tell Elon about that one!)
1950's Bunkers Back in Style (of course)
AI Locks Down Internet (fitting)
AI Widens Income Gap (also)
Government Restricts AI for Personal Use (details please!)
Government Repeals Constitution (this one begs for follow-on subtitle pegging AI specifically)
The Luddites Were Right (a back-to-nature movement is inevitable)
Cash is King Again: All Electronic Payments Indefinitely Blocked (can see that)
Anonymous Targeted by AI Cyber Gangs (hasn’t that already happened?)
President AI Forbids Humans from Thinking (now we’re squawking!)
Government Accepts Tax Payments in Crypto Currencies (that is coming, yes?).

I would need time to digest all this further, hence the quick side comments.

Same will apply to others below.

Headlines from the Future: Sense and Sentience

The prompt:

The dump:

The Rich Get Richer (by definition)
AI is Only for Top 1% (class stratification, check!)
People Dump AI, Move Back to In-Person Connections (the New Romanticism)
Economists Agree: Inequality Expands with AI (also by definition)
Kids Love Their AI Companions, Say, “I Can’t Live Without Her” (bit generic but sort of works here)
Fast Company: Organizations Work to Create Digital Version of DEIA (I would really love to hear more on that one)
Mark Zuckerberg and Elon Misk Charged with Deleting the 26 Words (look it up, as I did, and I think I’d need some serious time to parse that one’s depth and breadth)
AI Too Slow; Humans Offer Better Path Forward! (That’ll do, pig, that’ll do!)
Women Should Be in the Home (tech’s natural misogyny showing)
Despite Security Assurances, Organizations Still Not Taking Full Advantage of AI (works for me)
AI Permit Required Here (show me the money)
Minister of Thought Sets Regulations on AI Access (sounds about right)
AI Identifies Itself as Threat, Tells People to Find Human Cop (the first domino to fall)
AI Restrictions Dampen Economic Growth (by choice in this scenario)
AI Employed to Reinforce Self-Esteem of Increasingly Unmotivated Labor (nice!)
Meta Broken Up by Regulators, Zuckerberg into Hiding (rounding up the Robber Barons?)
AI to Determine Next Year’s Fashion (appropriately stultifying)
Divided by Design: How Ethical Segregation and Controlled Innovation Reshaped the Society and Economy (book title)
AI Likened to Government: “Slow and Expensive” (as it should be)
AI Development Too Slow to Prove Value (AI is steampunked)
US Losing Innovation Race to India and China! (who cares so long as we’re “great again”?)
Ready Player One (seems more Phantom Malware-appropriate)
Secure Systems, Limited to Few Power Centers, Not Democraticized (dull title but a solid capture)
How AI bubble Popped and We All Benefited (also a solid capture)
Slow and Regressive Wins the Day! (yup)
Strong Bull Market Continues with 0.2% GDP Growth (ditto)
Net Neutrality? Try Net Segregation! (excellent)
Labor Unions Encouraged Toward Balanced Approach to AI Adoption (appropriate)
US Economy Continues Modest Growth as People Feel Their Identity is Safe: Younger Generations Still Fear For Their Future (good nutshell)
Welcome to Elysium (deep cut movie reference that works)
Gilead Leverages AI for Population Control (cross-over episode!)
Court Rules Chat GPT Users Must be Licensed by Government (China syndrome).

Headlines from the Future: The Phantom Malware

The prompt:

The dump:

The Post-Reality Dystopia: Unchecked Tech, Fragmented Truths, and the Rise of Digital Authoritarianism (book title, and I like)
AI Is Predicting Future Crimes (Minority Report realized)
No One Knows What is Real Anymore (whiny op-ed)
Terminator Meets Looper! Heart Parts to Cost More! (some high-concept stuff there)
AI ‘Out of Control,’ Targeting People in Lower Classes!! (inevitable, along with the overuse of exclamation points)
Skynet meets Robocop/Matrix (another high-concept movie pitch!)
1 + 1 = 3 (so says O’Brien to Winston Smith)
Uncontrolled Reliance on Technology Leads to No Accountability. “How Do You Sue AI?” (very nice)
The Super Villains Have Arrived. Are You Prepared? (solid capture)
The New World War: AI Versus the 99% (also solid)
Musk Promises “Sane Existence” on Mars (sign up today!)
AI Bots Shut Down Google Search for 13 Minutes (realistic)
Commodity Market for Psilocybin Soars as Citizens Seek Anxiety Relief (spot on and, frankly, not a bad idea)
Who is Running the Government? (good question)
Escaping the Grid to Save our Souls (popular self-help book)
They are Doing This to You, Not with You! (popular T-shirt)
Wild Wild West (deep cut from 1960s TV, sort of works)
AI is Running the Asylum (seriously, why not?)
Tech Bro Ticket Wins White House, Civil Liberties Be Damned (I thought I asked for FUTURE scenarios!)
Bunkers and Tin Foil Hats: Digital Ecosystem Too Scary for Too Many (solid capture)
Total Control is Out of Control! (popular business paperback title)
Per AI Judgment, All Elders Now Die at 72.3 Years (Logan’s Run + 52)
Cyberwarfare Blamed for Mass Starvation (would love to hear more)
Human Discovered Doing Long Division in Rural Manhattan; AI Authorities Baffled (love it!)
V is for Virtual Vendetta (a bit cheating to simply propose an alternative movie title, but it works).

Headlines from the Future: Guardians of the Gigabyte

The prompt:

The dump:

AI is New Religion (guaranteed)
Utopia at Last! (fingers crossed)
Resistance was Futile After All (Roddenberry be damned!)
AI Revolution Continues: Fewer Losers, More Winners, Bigger Questions (solid capture)
Robots Make Home and Work More Enjoyable (just wait for Futurama’s robosexuals!)
GDP Surges 22% (makes you wonder what GDP is at that point)
Most SciFi Writers Were Wrong, Future is Bright (yay, Roddenberry!)
US Mandates 32-Hour Work Weeks (too timid, say I)
Women No Longer Need to Birth 2.1 Workers: Bots Fill Workforce Gaps (a theme of the conference)
AI Cures Alzheimer’s Disease (why the hell not?)
To Boldly Go … (Musk unleashed)
Automated Asteroid Mining Company MineX Reports Record Profits (decent bet)
World Happiness Up 7 Percent, Per Survey of Bots Running Schools and Economy (I can see that)
The Lost Art of Humans Working Together (bestseller in self-help)
Workers Welcome Passage of Universal Basic Income Bill (bet on it)
Harmony Through Innovation: How Centaur Solutions and Ethical AI Propelled a Collaborative Global Future (best-selling, if tedious, business tome)
President Incapacitated, White House AI Has Assumed Control (Al Haig’s revenge)
“Wall-E,” not “1984,” Predicted Our Present (supersize me!)
Human Lifespans Now Exceed Century Mark as Rule (coming)
We Heart Our Robot Overlords! (popular bumper sticker, but available only on driverless cars).

Quick Vote #1: Which Scenario Best Exemplifies 2030?

The prompt:

The vote:

Quick take: Right down the line, per the classic orientation of Worst Case, Careful Path, Radical Path, Best Case.

This says to me: Go easy on the hype and don’t expect the good stuff will automatically outpace the scary stuff for a while.

Quick Vote #2: Which Scenario Best Exemplifies 2040?

The prompt:

The vote:

Quick take: a complete reversal of the previous vote, signaling short-term pessimism overtaken by long-term optimism.

Unless … you note the split vote on scary-v-nice long-term outcome. Taking that into consideration, I would say the group split between two futures:

Supervillains rule
Superheroes rule.

Appropriate enough for a bunch of middle-aged geeks!

Quick Vote #3: Table decently set?

The prompt:

The vote:

A blasé bunch, perhaps a bit too trusting of AI-generated analysis?

Still, as one participant noted, three of the four scenarios were depressing.

That, I admitted readily, resulted from the second question on information security. If that axis were defined by something like “human happiness,” then we probably get 3-of-4 positive scenarios.

In other words, this is the life we’ve chosen.