What AI is Not
According to a recent survey, 43 percent of respondents in the U.S. and 46 percent of respondents in the UK answered that they have no idea what AI is about1. Of those who answered that they did have some idea of the definition of AI, according to Craig Martell’s lecture, they are likely incorrect. There is a pervasiveness in everyday conversation around the use and promise of AI. The approach is equated to human cognition with the potential to change every facet of our life, to make daily processes easier and to eventually replace the need for the human workforce.
Like most concepts that take over the collective consciousness of groups who likely do not have the foundational knowledge to understand the underlying components, the implications of AI are both anticipated as well as feared. The confusion is justified. When paired with “artificial”, the word, “intelligence” can create all manner of expectations fit for a sci-fi film. The definition of the word, “intelligence”, however, is diluted when referred to in this way. Pattern recognition based on repeated tasks and labeling yields to enhanced determination and prediction of an object category. This does not lead to the capability of broader inferential conclusions nor does it mean that the system is “learning” beyond the best algorithm on which the pattern recognition is based.
The conflation of terms like artificial intelligence, machine learning and neural networks, however, has spawned an industry that markets these capabilities to sell into companies and industries where knowledge of these concepts is lacking. At the same time, salaries for top AI researchers have skyrocketed because there are not many people who understand the technology and thousands of companies want to work with it2. They hype is a disservice to the field. Progress requires patience and the promise of the technology has been made so hyperbolic that there is an expectation of eminent and transformative results. The education that individuals like Craig Martell are providing is necessary. There needs to be a campaign to correct the misinformation, set expectations and buckle down for the long road ahead.
- Renee Buest, Artificial Intelligence is Not a Threat to Human Society, IDG Contributor Network, Feb 2018.
- Cade Metz, I. Researchers Are Making More Than $1 Million, Even at a Nonprofit, NY Times, Apr 2018.
Users who have LIKED this post:
8 comments on “What AI is Not”
Comments are closed.
I’d be interested in hearing more about why you think the hype around AI is a disservice to the field. Does it create unrealistic expectations? Is it hurting research efforts?
I think overall people know that Technology has taken over a lot of their most basics activities of their day to day however they hardly ever think about what is behind and actually just think of it as a given. The amount of people that are touched by AI nowadays is huge compared to what our parents and previous generations had, even my grandma facetimes with us and knows how to use spotify!!
The question is how quick the different education levels will evolve into a curriculum that not only helps the society understand what it is, how it works and the benefits and threats but also how to develop it. A lot of good programs have emerged to get early involvement from kids however we’re still ways away.
This is a thoughtful piece, mona2. Thanks for sharing.
It reminds me of the AI Winter of the late 80s and early 90s. Back then, the hype was certainly damaging to the field, in its own way, by propping up expectations (and valuation of “AI companies”) so high that reality was doomed to fall short.
I would defend, though, the “thousands of companies” that want to work with AI — having been in the field for a while, my overall impression is that AI is what they *say* they want to work with, but to get them to actually write a check (for software or services) and stick with a real project plan… Many (most) are not there yet.
Unfortunately, it’s just like selling software, hype is always involved (DOMO anyone..?), and the buyers sometimes want to be associated with the buzz too… (market is all about signaling. sigh.)
🙂
Thanks for the post!
I think to myself after this is that if AI doesn’t really give a machine a mind of its own then why would people need to fear them.
There seems no possibility of machines being capable of waging war or going on a killing spree.
In such a world we dont need to fear new technologies
rather embarace them and use them for ease and comfort
It is great to think deeper about what AI is not. That includes our understanding of the terms intelligence and neural networks. We learned that these terms were and are used in large parts to generate funding for related research. The advancements in these fields have been impressive, and advancements in science and technology are always positive, I would argue. However, intelligence includes self-awareness and reflection. The great thinkers and brains of humanity always searched for ways of understanding and explanations of items beyond the current scope of thinking. Diving deeper into this would give us a better understanding of how close artificial intelligence is to human intelligence since this goes beyond drawing inferences from impressively large datasets.
It is true that nowadays people are consumed by the hypes around AI, and I think it is great to think deeper into what AI is not. But I would ask a question differently: in what sense do you think AI is different from human intelligence, and if they are indeed different, then how could we explain human intelligence? Craig Martell compares machine learning with a “input, black box, output” model, and in which sense does that different from human analysis process? I know it sounds humans are more than just that, but are we actually?
Before listening to the lecture about AI, I had a wrong definition of AI as many people do. After the lecture, I realized AI is only a popular buzzword. It is important to let people know what AI really is. This will dissipate concerns and fears.
I believe that if AI is not really understood is because people may think that is something of the future and that does not apply to their day to day life. In fact, AI is something thats being going on since long time ago by now. The ability to be able trough a number of inputs as sensors, determine in a logic way the actions to take, is something that has been taking place in our world quite a few decades yet. The more this concept will be used and developed by companies, the more people will need to understand what is exactly AI and how to use it.