Whats relevant? Whats not?
“As you ramble through Life, Brother,
Whatever be your goal.
Keep your eye upon the doughnut,
And not upon the hole.“
13 January 1929, Charleston (WV) Daily Mail, pg. 6, col. 6:
Sign in Quarrier street restaurant. Advice for those who drink coffee and eat “sinkers.”
The original AI Winter
Artificial Intelligence is not new. Another thing that is not new is people over-promising and under-delivering results around AI. In fact AI in the modern era (I won’t even go back beyond the 1950’s but if your curious there is a great article here) has suffered a number of funding declines as the market has realized that that particular version of AI’s promises were not going to pan out.
After the initial heyday of AI from the 50’s until the mid 70’s substantial research dollars were squandered. As companies and government agencies realized they had been hoodwinked on the immediate value, they pulled dollars out of those investments. This AI winter lasted pretty much up until the early 80’s.
The agencies which funded AI research (such as the British government, DARPA and NRC) became frustrated with the lack of progress and eventually cut off almost all funding for undirected research into AI. The pattern began as early as 1966 when the ALPAC report appeared criticizing machine translation efforts. After spending 20 million dollars, the NRC ended all support. In 1973, the Lighthill report on the state of AI research in England criticized the utter failure of AI to achieve its “grandiose objectives” and led to the dismantling of AI research in that country. (The report specifically mentioned the combinatorial explosion problem as a reason for AI’s failings.) DARPA was deeply disappointed with researchers working on the Speech Understanding Research program at CMU and canceled an annual grant of three million dollars. By 1974, funding for AI projects was hard to find.
The second wave of hype started in the early 80’s and went through the early 90’s. This wave included Expert Systems, Logic Programming and LISP Machines (anyone remember Symbolics).
Again, a huge amount of over promising produced a hype wave – that ultimately lead to a ‘market correction’. I had personal experience with this one.
As a young student, with recently minted PROLOG skills, I was eager to jump into AI in the late 80’s I was sorely disappointed when to my surprise a lot of the AI companies just went away by the time I was hitting the job market. I ended up writing code in the CASE tool business for most of the 90’s until the Internet became the new sexy toy. All during this time – AI was mostly no-where to be seen. Over 300 AI companies went out of business by 1992.
Jumping the Shark Again?
Why with all the past evidence of over hyping AI do companies still fall into this behavior? Simple. Its the money stupid.
“Recent research by ‘Price Intelligently’ shows a 30% increase in purchase intent from companies to buy products that have an AI integration in product offerings.”
– Jiaqi Pan
But people are only so willing to pay for poor results. When they are disappointed again and again – well – eventually they go spend their money elsewhere. And that’s pretty much what has been happening. I literally don’t have the time to provide a comprehensive piece dedicated to AI failures over the last couple of years… lets just touch on a few.
IBM Watson for Oncology
“This product is a piece of s***. We bought it for marketing and with hopes that you would achieve the vision. We can’t use it for most cases,” a doctor at Florida’s Jupiter Hospital was quoted as saying in the documents, according to STAT.
-Customer feedback from internal IBM documents (while at the exact same time IBM touted major success to the press)
Everyone has heard of Watson. Its hard not to with all the marketing dollars IBM spends on it. But they failed to deliver on the very specific narrowly focused goal they set out to achieve. This was not “General AI”. This was just pattern matching on X-Rays using Machine Learning.
…the healthcare news publication Stat published a report claiming “internal IBM documents” showed the Watson supercomputer often spit out erroneous cancer treatment advice and that company medical specialists and customers identified “multiple examples of unsafe and incorrect treatment recommendations,” even as IBM was promoting its AI technology.
This resulted in a major shakeup at IBM. People lost jobs and funds we reallocated. While the CEO of IBM still firmly maintains that they did not hype the results – fewer and fewer people are listening to her.
Seven Dreamers Laboratories dreams no more
Seven dreamers laboratories was a Japanese company that specialized in the development of AI, robotics, and healthcare devices. It raised over $104M to build diverse products like a machine that would fold clothes, but also medical devices that aid breathing and 3D measurement devices. With regard to their last run (the folding machine)
Clearly, the product was too ambitious. Seven Dreamers had planned a simpler, but still potentially-impressive version that merely folded and sorted clothes. The first-gen model still required a complex combination of robotics, image analysis and artificial intelligence to achieve its goals, however.
The company folded yes (ok – that was funny) but not before wasting an enormous amount of investor moneys. Remember that. I can promise you investors do.
Uber self-driving car kills a pedestrian
Not much really needs to be said on this one. Yes… she was probably jaywalking. Yes the controller was playing video games. But someone died. At the hands of a AI driven car.
Uber’ self-driving software declined to take action even after the car’s sensors detected the pedestrian. US National Transportation Safety Board’s preliminary report on the accident made it clear that the software failed.
We may eventually get self driving cars – but it won’t be in the way that is currently envisioned. More on that in a future blog.
I could go on forever. Go search “AI Failures” and you will have a lot to read. From Amazon’s AI recruiting tool being a sexist pig to Microsoft’s Twitter bot becoming a racist, misogynistic anti-semite. AI failures are the norm. And it’s not just me noticing.
The Investment community takes note
Perhaps more disconcerting is the investment community’s reaction to the current state of the AI market. They don’t like it. Not one bit. Below is a graph of deals done with AI companies over the last 8 years. As you can see – 2019 is set to be a major drawdown in terms of investment into AI by the VC communities. While its only July – its clear that the VC community views AI as having Jumped the Shark.
2019 is looking like a major step back for AI.
One thing about VC’s. They put their money where they think they will get a return – and they are generally pretty good about getting it right over time. They tend to watch the donut and not the hole.
The organization behind TrustNoRobot is now a non-profit! Our mission is to promote humans above technology. Our simple belief is that technology should work for humans and better their condition and never have a negative impact on humanity. Please find it in your heart to donate to this important cause.