Four challenges in gaming and learning innovation with AI

In many ways, it seems that with AI in business, the sky is the limit. Countless innovations convert industries from the ground and from the top to bottom. Consumer technology and business technology are tall towers that are constantly enhanced, so often to meet in the middle.

But the challenges remain. Here are some of the usual issues that leadership teams have to face in the development of innovative AI applications.

Quick innovation and quality bar

A principle of the development of AI brings to mind that the old saying that the execution of a project in integration is often “easier to say than to be done”.

Or it might be better to change this slightly: “It’s easier than the sophisticated.”

The idea is that with the democratization of the code and all these other trends, a non-technology person can create an application quickly-perhaps even in a few minutes.

But supporting this application, ensuring that it is without errors and works as intended, is much more difficult. There we come across some of these “tooth borders” of AI capabilities. He can do things, but he can’t always try things so well.

The higher the quality line, the harder it is to support and maintain a project that may have started very easily (see the excerpt near the end of this post.)

Repeat kills the number of users

The second challenge, which is common in gamified and learning applications, can also be summarized with an old saying, this time in the form of a question: What did you do for me recently?

Or you could put it differently by saying that users tend to be unstable and demanding in the long run.

Here’s an example of how this idea could hurt an application – let’s say that the first repetition of the project is simply amazing for a large basis of users and many people sign, wanting to use these things to their advantage.

However, in addition to using successive software versions, they find that they are buggy or glitchy, or does not work as intended or advertised.

And the project begins to bleed users.

This can happen in the beta phase, or even later. Again, it may have to do with rapid growth and lack of overall support. It is the idea that a central condition for AI may be amazing, but if it is not well designed, it may not deliver what users want and so they end up leaving the ship over time.

The seller’s bugbear

If you’ve ever heard a company say that “AI accounts are high”, what you are dealing with with a business trying to offer either a wrapper or a third -party service based on someone else’s models.

This means that they have to buy AI functionality to run their platform and can end up paying more than they want.

After all, this is only the cost of business, but it can be a problem for a development team if they do not have their own internal systems or are unable to work around the supplier’s limits.

Building without code, features and on the spot

Again, you can create something easily, but you may not have all the features or functionality you need.

Take the role of the boat. Suppose software uses AI to produce excellent user results, but it is so difficult to connect with users, again, are running to shout away.

More about AI projects for learning and gambling

A recent fantasy in the Stanford Action Department has gone through some of these issues, where leaders in language and learning companies talked about some of their experiences.

Bing Gordon of Kleiner Perkins interview with Natalie’s look at Duolingo and Kylan Gibbs of Inworld.ai on “Building Smart Experiences”. The two covered the challenges mentioned above, which I will point out with some excerpts. They also talked about goals such as the development of talent.

“We have a talent acquisition machine and a boarding machine that has been happening for years and works,” Glance said. “It works very well. We do our job and get 1000s of applicants in a few days, so we get our choice from the best and brightest in all the United States, around the world and so it’s great.”

It analyzes this process a little, describing a virtuous circle of users involved in the business.

“We get people who are really, really good, who are really excited to work for Duolingo,” he added. “They may have used it from high school and start with them when they are still at university. So they come, usually as growing young or rising elderly. We have a practical program so that they really learn about work and prove themselves and at work.”

Innovation with NPCS

“Basically we built (living tools of interaction) as a way to build mainly game characters, NPCs, you know, learning applications, these things,” Gibbs said. “And a lot of what we did was to build the immediate structure around it, so that we could have these interactions in real time to put personality and then what happened to Chatgpt, I think, is that people have decided or decide that the factor of form we all use …

This, he said, produced a deliberate response to his company.

“We just moved down (inside) the stack,” he said. “We focused more on speech models, language recognition models, as well as, you know, things like knowledge and then found that the way we first built it was this kind of single box where everyone can build a character in the same way, but the different characters have different characters.

Using a basis of knowledge

Glance and Gibbs also talked about the use of memory, which Gibbs was described as a “basis of knowledge”.

“Then you will decide. How will I use them effectively during the conversation?” Said Gibbs. “Will I really bring them all the time? … If you talk to someone and they remember every thing you said, it will be really weird, and especially if we repeat it back to you and so I think it’s not only: how you store this memory, but – how you process it, and then, a theme of conceptual classification.

More excerpts from the table

“Most people you are talking about, including some C -level executives, are like:” Well, I built this thing on a weekend and it’s very awesome. – Kylan Gibbs, in Dev circles

“Our main priority is to create a truly, really good feature. This is what comes first and then we take care of the second for the speed of the product and only the third for the cost. It proves that the perfection is quite fragile. – Natalie look, for work at cost

“Everyone has their own world that they are coding.” – Kylan Gibbs, for democratization of coding

The world of the businessman

I hope some of them are useful to show leaders what to do in AI’s playground. For more information, watch out for the blog. Stay tuned.