This article is more than 1 year old

Is today's AI yesterday's software routines with better PR? We argued over it, you voted on it. And the winner is...

Spoiler alert: Not that machine-learning salesperson

Register Debate How does that saying go? I'm not a cynic, I'm a realist. That's pretty much how I'd sum up our first-ever Register Debate, which ran this week.

Over the past few days, we pitted some of our vultures against each other, cajoled our readers into chiming in with their own comments, and took your votes on whose side you were on. The motion up for debate was: Artificial intelligence in the enterprise is just yesterday's dumb algorithms rebranded as AI. There is plenty of evidence that software, equipment, and internal projects are being labeled as powered by buzzword du jour AI, and buyers are spending millions on them.

The "adoption of artificial intelligence is growing worldwide," IDC declared in June. "Over a quarter of all AI initiatives are already in production and more than one third are in advanced development stages. And organizations are reporting an increase in their AI spending this year."

But how much of it is actually artificial intelligence? Is today's intelligent tech just the same old algorithms from previous versions with better marketing, or is real innovation happening? And if proper machine learning really is being deployed, what are the business advantages given ML's strengths and weaknesses?

Kicking things off on Monday, Thomas Claburn argued in favor of the motion. He noted that AI was so slippery to define and amounted to modern-day alchemy.

Most companies looking for data scientists are looking for people to collect, manage, and calculate basic statistics over normal business processes

"The vast majority of businesses are still in the early phases of collecting and using data," Carnegie Mellon University assistant professor Jeffrey Bigham told our vulture. "Most companies looking for data scientists are looking for people to collect, manage, and calculate basic statistics over normal business processes."

Svetlana Sicula, a research veep at Gartner, told our Thomas that there's still a major gap between the science of AI and the engineering of AI. For example, the COVID-19 pandemic broke many AI solutions for fraud detection and supply chain management because as people's habits suddenly changed, the incoming data suddenly changed, and the models couldn't handle scenarios they weren't trained for. As such, whatever's in today's enterprise AI may not actually be all that intelligent or new or practical, judging from these sorts of breakdowns.

Next up was Katyanna Quach who wasn't going to let cynicism triumph, arguing against the motion. She pointed to cloud-hosted machine-learning services offered by the likes of Amazon, Microsoft, and Google that can be harnessed by today's business applications to analyze data in new ways – sound and object recognition, pattern recognition in customer records, and so on. These are solving real problems for big names in retail, energy, and science.

And manufacturing, too. "Even though we are just getting started, the systems we have deployed in Europe and North America are already learning from one another and improving every day," she quoted Pieter Abbeel, founder and chief scientist at startup Covariant, as saying regarding the AI-powered factory robots it makes and sells.

And healthcare: computer-vision models on the market that can detect damage to people's eyes. When you get an AI-based product from a trusted supplier, you really are playing with some level of artificial intelligence.

Top reader comments as upvoted by you and selected by us

"Without doubt there are machine-learning algorithms that are extremely useful, particularly in industry and medicine. Are they clever? Yes. Are they intelligent? No!" Chris G

"Statistics is about understanding the relationship between input variables. Artificial Intelligence (be it evolutionary computing, neural networks, learning classifier systems or a mix thereof) are about matching input to output. For example: a neural network cannot tell you what the relationship is between your height and weight but it can guess your BMI.

"Machine learning algorithms aren't that new but it's not statistics." RobLang

"The reality is that AI is still somebody's algorithm, only justified with mountains of data.

"Until there is actual intelligence - i.e. independent reasoning combined with intuition and reinforced with scientific method, "AI" is pure marketing garbage albeit very useful for killing jobs for people." – c1ue

After that dose of reality, Dave Cartwright poured a bucket of cold water all over the motion, also arguing against it. He took a rather cunning angle, that the motion has to be wrong – because yesterday's algorithms can't do what today's AI can do, so no, artificial intelligence isn't yesterday's software routines.

Machine learning today involves mountains of data in all sorts of formats that have to be processed into a form that can be used to train and test neural networks and similar systems, and then these models have to be able to make decisions on new data flowing into them. All of this goes above and beyond what your typical heuristic functions and binary decision trees are capable of. Thus, there must be some intelligence present, judging from the work being carried out.

Like cavalry coming to Thomas's aid, Rupert Goodwins closed off the debate, arguing for the motion. Well, not so much arguing against it as performing a perfect drop kick to it.

Take a technology few in business fully understand ... then sell it as the most important new thing any company can have. Don’t quite know what’s going on? Perfect, you won’t be able to tell whether it’s working or not. Sign here.

An artificial intelligent algorithm sounds technical and mysterious, he opened with, adding: "The use of the exotic to mask the mundane is indeed diagnostic of what’s going on.

"Take a technology few in business fully understand, define it as needing lots of expensive everythings - hardware, software, people, bought or hired - and don’t say exactly what it’s doing or how it’s doing it.

"Then sell it as the most important new thing any company can have, to fend off all the competition already adopting the new shiny to reduce you to rubble. Don’t quite know what’s going on? Perfect, you won’t be able to tell whether it’s working or not. Sign here."

Rupert went on to argue that enterprise AI is stats, it's analytics, and it takes you away from the decision-making process if you let it boss you around. And you're never quite sure if you can trust any surprises in the insight offered because no one seems to be able to explain how it came to that conclusion. Enterprise IT, on the other hand, is all about being able to understand the output, and if the output is unexpected from the input, you know exactly how to fix it, with the real intelligence taking place between your two ears.

So, who had the best argument? Who won? I'm not going to call it: I'll leave that to the reader vote, in which no ballots went missing in the mail. The final score was 69 per cent for the motion, and 31 per cent against. Thus, the majority of you who took part reckon today's AI algorithms are yesterday's software routines with better PR.

What's my takeaway from all of this? A functional definition of AI: a system that's intelligent until it isn't. ®

More about


Send us news

Other stories you might like