What is Artificial Intelligence? Will the funding of AI create the Terminator?
Based on some recent news articles ranging from the NY Times and warnings of super intelligence to Elon Musk and the Terminator, there is definitely a threat that technology can go too far. Certainly, any topic that has the super-tech-optimistic Elon Musk worried about the downside of technology should alert some concern. (Of note, he is clearly much more knowledgeable then me on everything.)
But, Elon Musk is still investing in AI. Perhaps, as noted, it is simply to keep “an eye” on the technology.
Why? If we take the baseline assumption that people do not like to lose money, then there is likely a large opportunity in the future AI.
Slate presents the other side of the coin, why not to be afraid of AI, with the perspective that most people only have a vague idea of how technology works and trust the system.
Instead, we are forced to trust that the systems and subsystems we depend on and the experts who maintain them function as advertised.
Personally, I have no idea how bridges over large bodies of water were ever made. I can dig a hole more than a foot without water seeping in, so how was the Brooklyn Bridge constructed in 1883?
There are obviously pros and cons of AI, so what is causing the advancement in AI?
From Wired, the future of AI revolves around three breakthroughs:
- Cheap Parallel Computation
- Big Data
- Better Algorithms
Essentially, the AI revolution is based on cheap computing gathering massive amounts of information to teach a machine how to learn.
Obviously, the VC community has noticed and invested $17 billion since 2009, including $2 billion in 2013 alone (in 322 companies). The AI investments represent a 62% growth in the last 4 years. CB Insights produced a list of the top AI firms to watch. The list of firms, which are backed by the who’s who of VCs, mainly focus on machine learning and cloud services. Machine learning points to teaching context.
So what does AI represent? Does it represent the next big thing?
It appears to represent the path to solve discovery. If the Google search algorithm can tell the difference between a search for Main Street, Anytown USA and the “main street, Anytown USA”, then there are signs of the discovery use case being solved. An ambiguous search term provides the correct results in context.
In fact, companies such as Yahoo!, Dropbox, LinkedIn and Pinterest have been among the AI investors. Would those companies benefit if they could produce the discovery algorithm that matched the Google search algorithm?
Within the context of discovery, but more meaningful, then will AI add value to previously labor intensive tasks? Although, I’m not involved in law enforcement, AI could have theoretically helped in the Boston Marathon bombing investigation.
Law enforcement officials are trained diligently into finding minutia, the needle in the haystack. Officials could have worked with technologists to create a crisis response AI program. The advancement of machine learning would be trained to look for non-responsive faces or movement away from the explosion, which took the good ole man power to scour the tapes following the bombing (see the 60 Minutes report following the arrests). Is it possible that the suspects could have been identified earlier? I think so.
Would that have saved the MIT police officer and Watertown lockdown?
Would that have prevented the inaccurate Reddit sleuthing?
Would that have stopped the mistaken capture in Watertown?
What does this present? AI provides a scenario where the technology is overstated in the short-term, but understated in the long-term. Additionally, there are perhaps two states of AI. The consumer focused enhancement of the “Amazon recommendation” that leads to proactive search and the government/military grade technology that leads to Terminator nervousness.
O’Reilly Radar presents the possibility of AI in context of human intelligence.
We don’t understand how human intelligence works at a fundamental level.
If we don’t understand how to think, then how can programmers simulate human thought into machines?
But what do researchers and programmers know? They know how humans make decisions. They know how to provide context, which can assist humans.
This leads to opportunity. Technology enters the market at the lowest level or ease of use. In this case, AI technology can adopt to the consumer market and make valuable improvements that assist people. As the underlying technology advances and developers can implement safeguards (under the watchful eye of Mr. Musk), then there is a conceivable future of AI on the landscape.
From the same Wired piece describing the future of AI:
As AIs develop, we might have to engineer ways to prevent consciousness in them – our most premium AI services will be advertised as consciousness-free.
Will it play out like this or anything close? Absolutely not.
The markets will back the companies that develop solutions for the mass audience. The best discovery advancements will control the market. The companies that focus on military advancements will be worrisome (but how much does the average person worry about military contractors now?).
Ultimately, why is this important?
AI presents an opportunity to redefine a market. AI very well may create a new market with new economic incentives, which has a massive impact on wealth creation. Investors back and support the development of moonshot opportunities (at least in theory).
If consumption is about 70% of US GDP and retail sales were $4.5 trillion with $40 billion in ecommerce (see Emarketer), then investments flow downhill. The low hanging fruit of AI is aimed at the consumer and influencing the purchase process.
Can AI deployed correctly impact the consumer? Yes. If I had an opportunity to invest in a talented team targeting this market, would I invest? Yes.