Is AlphaGo Really AI? What AI Strategy Should Your Startup Adopt?

Screenshot courtesy of Stanford CoreNLP — Core natural language software

The first 6-game match between Gary Kasparov and Deep Blue, played in 1996, ended in a 4–2 win for Kasparov. The world slept better after those games. But our joy at the superiority we still held over ‘the computer’ was short-lived because just over a year later, Deep Blue defeated Kasparov 3 1/2 — 2 1/2 games. It seemed our subordination to the machines was inevitable.

But our robot subjugation never came. What we got in 1996/1997, due to the advancement in technology, was

• Wireless internet: while the decision to open up the spectrum came in 1985, it was only in 1997 “ the committee agreed on a basic Wi-Fi specification. It allowed for a data-transfer rate of two megabits per second, using either of two spread-spectrum technologies, frequency hopping or direct-sequence transmission” standard did things really pick up for wireless internet technology we take for granted nowaday.
• Virtual reality therapy: Fully altered sensory perception therapy started in 1997.
• Dolly the sheep: the clone that shook our world
• Thrust SSC: the jet-propelled supercar that till this day holds the land speed record.

These inventions were all “amazing” and “the next great thing” and other superlatives but they aren’t as amazing now in the world of CRISPr, self-driving cars and Machine Learning. Some of these scared us, remember the furore over Dolly, but we’re still alive developing technology to solve the problems we have. Some of them scare us like AI does today. But we survice, because that’s what we do when our technology advances; we continue to solve our problems.

AI and What Scares Us Today

I recently worked with a software developer who believes that it is inevitable that Artificial Intelligence will ‘take over’. Like every zealot before him, who has believed in the inevitability of some event in their religion, he thinks this time is THE time. He also believes only the smart people like him will survive and thrive. But I have to disagree with him as I share below.

Some argue that AlphaGo is not AI and some suggest that the technology giants (and their investors) are hyping up the technology that most of them are claiming is at the center of their strategic futures (Google, Facebook, Amazon and Microsoft). I’m in the camp of folk who believe we are advancing greatly with our technology but, paraphrasing Dan Ariely,

“AI right now is like teenage sex; everyone talks about it, nobody really knows how to do it, everyone thinks everyone else is doing it, so everyone claims they are doing it…”

While we are as advanced as we’ve ever been with machine learning and neural networks, we are only just at the very tip of the very edge of the beginning of the cusp of the age of Artificial General Intelligence (computers as smart as humans across the board). AlphaGo winning Ke Jie does not mean AlphaGo is smarter than Ke Jie. It just means AlphaGo can win Ke Jie at Go. AlphaGo cannot feel and emote like Ke Jie. We are a long way from the days when our robots become more intelligent and more human than we are…and it’s that humanity that will be the winning strategy for utilizing AI to serve our needs.

What’s Your AI Strategy?

According to Creative Confidence by the Kelley brothers, with every innovation is the need for viability, feasibility and desirability. With Artificial Intelligence

• We are learning that the technology is feasible (technical)
• We are getting close to viability for many use cases of artificial intelligence but
• Where we are failing is in getting it to be attainable and desirable to all business people.

Because something is feasible and viable doesn’t make its success inevitable. this is because of the human element of desirability. The narrative of Artificial Intelligence right now is one that scares the rest of the world who do not quite understand what it is and what the impact could be. Especially small businesses that do not have the capacity to develop systems of their own. Thankfully, Google (with Tensorflow) and Amazon (with AML) are open sourcing technology for all who desire. The next step would be for some entrepreneurial developers to build products to serve small and medium size businesses.

Tensorflow.org

What if you’re a small business and you’re wondering how to use machine learning/AI? What if you want to “use intelligent software agents that receive percepts from the environment and take actions that affect that environment” in your business? The best questions to ask to center around

1. Do you want to use AI to assist (improve what you currently do)? This approach works well for businesses that are looking to increase efficiencies or reduce the cost of their current methods of doing business. Businesses can think of this use of AI more as an internal use case where the customers might not know it’s AI but it will improve how the company does business.
2. Do you want to use AI to augment your business (do what you couldn’t previously do)? Is your business in need of new technology? This use case for AI is more geared towards developing a new product/technology to serve the customer. It can be thought of as external facing use cases that helps the company solve customer problems innovatively.
3. or do you want to use AI to automate your business (create machines that think and act on their own)? this is the breakthrough strategic approach and is the tougher one for most small businesses. Considering the work and resources the larger companies are pouring into this, it might be best for small businesses to utilize some of the open source platforms mentioned above.

Nothing says you cannot mix and match these approaches. Key is to focus on the most beneficial goal and let that drive your approach to finding partners who can help you build a solution that best serves your business need.

Ke Jie lost to AlphaGo and the post-mortem will happen for the next few days/weeks. Everyone will claim the end of humanity is near. But I’ll suggest it’s just time for us to shift our attention away from how much better AI is than humans, and start doing the work (at scale) to ensure that humans become better with AI. It’s time for the technologists building these great tools to (for example) go on a 24hr drive with the truck driver to ensure that the self-driving car software we develop augments the truck driver in some way. It’s time for developers to, before even writing one line of code, sit with the telephone call center employee and live his/her life for a day instead of building software that totally replaces that call center employee. This approach will bring in some empathy and change the products we build.

Otherwise, as business people, we’ll continue to let the premise of the technology scare us from gaining the benefits we can from them. And as technologists, sitting on the other side of the business people, we’d just have transferred our dumbness to our AI.

And that would be pretty (uhm) dumb.

If you liked this you should check out ’10 Startup Lessons from 40 Books I read in 2017' and ‘9 Books To Boost Your Understanding of Technology Systems’.

Please like, tweet, share and heart the article if you enjoyed reading it. Sign up for my Polymathic Monthly Newsletter here, you’ll love it. Also, check out HarperJacobs for compelling content creation & Asha Labs for innovation/strategy consulting.