Original article was published by Oscar GR on Artificial Intelligence on Medium
What conclusions do we get?
The quintessential product. In my experience, it is the most widespread Use Case among companies, although in most of them I have not seen a clear return.
According to Gartner, in the next 2 to 5 years, chatbots are expected to experience more than 100% increase in their adoption rates, with a penetration rate that has risen from 5–20% in 2019, to 20–50 % this year.
Main suppliers? AWS, Cognigy or Google.
Artificial General Intelligence (AGI)
Or the ability for a machine to be able to understand the world as well as the human being, acquiring the ability to learn to carry out a large number of tasks.
While the Specialized AI is just that, specialized, being good at a specific thing like recognizing photos or recommending the next TV series that you will see. That is where the difference lies from General AI.
General AI is great at any task (no excuses!).
Specialized AI is created to do one thing, General AI is created to learn to do anything.
However, this technology currently lacks commercial viability, causing companies to focus on much more specific Use Cases that add value.
And now, the new entrants…
1. Small Data, First New Category
Gartner defines it as the set of techniques that allow to take advantage of much more resilient models where the starting point is found in much smaller datasets, thus allowing a better adaptation to unexpected events and where machine learning must be balanced with human domain expertise.
Transfer Learning comes into play, or the ability to transfer knowledge gained from one task to the learning of new tasks without the need to feed the machine with massive new sets of data.
2. Generative AI, Second New Category
According to MIT, “one of the most promising advances in the world of AI in the past decade”.
What does it consist of? Generative AI is based on the ability of computers to learn the pattern that exists underlying the input offered and on the subsequent use of that pattern to generate similar content, like text, audio files or images.
The so-called GANs, or Generative Adversarial Networks, are the generative models that use two neural networks that face each other: a generator, in charge of generating new data or content that resembles the original; and a discriminator or discriminative network, which is responsible for differentiating between the source data and the generated data.
3. Composite AI, Third New Category
Described as the combination of different techniques that improve the way of learning and acquire “common sense” to become AGI, Artificial General Intelligence.
4. Responsible AI, Fourth New Category
An umbrella strategic term that collects all aspects of commercial and ethical decision-making that organizations must adopt or follow correctly, including aspects such as business and social value, transparency, fairness, explicability, security or privacy, among many others.
5. Things as Customers, Fifth New Category
Gartner defines Things as Customers to any device that obtains services or goods in exchange for a specified payment.
I find this block very exciting because new opportunities will be opened to discover by having to deal with services instead of people.
That is, in what way do we promote a product or a service to a thing?
Yes, you read it right!
How do I make that device that belongs to a human, a family or another company decide to contract my service on behalf of its owner. How does a company appeal to a thing without emotions?
We will have to deal with what Gartner considers as Thing marketing, or other forms such as Thing service or Thing commerce.