Artificial intelligence – long way to go from hype to hope
ARTIFICIAL intelligence is here and it’s present, but despite the many millions of words that have been penned on the topic, the space between the imaginative potential of AI and the hollowness of its unfulfilled promise has not garnered enough attention. Indeed, “artificial intelligence” has become a buzzword, which has had the effect of masking the reality of its existence.
— sam scarypadi 🌶✊🏼 (@sam_sicilipadi) June 28, 2017
The promise of AI is also deeply encoded with cultural tropes such as those found in films like Bladerunner and The Terminator series which has had a significant hand in informing how everybody thinks about the potential of AI to disrupt our everyday lives. Just how much of our conceptions of AI is rooted in actual research, and how can we discern between what’s hollow and what’s real?
However, despite its potential to transform our processes and drive our businesses, AI, as we conceive of it today, is largely a product of hype. It can do many things, but many businesses today think of the technology as a kind of panacea to all their problems.
That’s not to say AI is not useful – the technology is already deeply embedded in our culture and frameworks, largely due to the confluence of these factors: the availability of hardware, the explosion of data science, and the improvement of our software and algorithms.
These factors have allowed researchers to tread closer to the heart of AI than ever before. It has yielded some pretty impressive technologies with the potential to become a “revolutionary gamechanger”, according to Lim Swee Kiat of startup KeyReply, who also moderated a panel at Echelon Asia Summit 2017 on the topic.
The panel was made up of three people working in the AI space – Annabelle Kwok, the CEO of a hardware-focused AI startup SmartCow, Jarrold Ong, the CTO and co-founder of SWAT, a startup working on providing urban mobility solutions, and William Klipgen, a managing partner at Cocoon Capital. They spoke about just how much of what we think of AI is pure hype and how much is hope.
— ✨Casie Lane✨ (@CasieLane) June 28, 2017
“I think it’s more hope than hype, but at this point, there has been no real breakthrough and it’s not always going to be a silver bullet,” Ong said. He made the point for his company, AI’s benefits reside in the potential of data to help us make better choices and drive the productivity of our businesses.
Klipgen concurred and added various verticals can be improved using AI.
“It’s hard not to see businesses implementing some kind of AI,” he said, noting the technology in its simplest form can deliver a lot of value to companies willing to engage with it.
Ong’s point about AI’s inability to be a “silver bullet” is key here – AI isn’t valuable because it can magically transform a business, but it has the ability to become a powerful multiplier that could help entrepreneurs become more productive when it’s implemented where it’s most needed.
Kwok related a meeting she had with a leading security company, where executives wanted SmartCow to replicate Amazon’s mammoth checkout system. She noted companies had an unrealistic idea of what leaps and bounds technology today had made and the hype around AI had translated poorly communicated vision into outsized expectation.
How exactly did this outsized expectation get created though? Kwok explained the two “AI winters” in the 40s and the 60s came about because over-hyping of the technology led to an outpouring of capital into the space, which in turn led entrepreneurs to overpromise. Of course, the technology to achieve what we have today didn’t exist back then, thus spooking investors who quickly dropped many developers, plunging AI back into obscurity.
The object lesson here is to be wary of the power of hype and how it has the very real power to kill a budding technology space with lots of potential.
“We need to be very careful with managing the hype and the hope when we talk about AI,” Kwok said.
“People have a very conceptual idea of what AI can do, but like any technology, it has limitations.”
Furthermore, since AI today is a largely malleable space, many companies are still unfamiliar with precisely what developers can train their technologies to do. Kwok’s experience with the security firm is a clear example of the lack of information and due diligence on the part of clients to educate themselves on what capabilities are available to them. The lack of information means clients largely rely on hype to guide them.
“It’s very important to communicate the vision [clearly] because it’s the stories people tell today that end up driving the technology we make,” Kwok said.
For the record, Kwok doesn’t believe another AI “winter” is imminent because many large corporations such as the Alphabets and Baidus of the world know what they’re doing in the space. But there definitely needs to be a bigger conversation between all parties involved in the scene in order to drive AI further away from hype and closer to hope.