AI in Education – A Realistic Look at the Effectiveness of AI in the Education Sector

A realistic view of the current adoption rate of AI in education, and pointers on how to ensure that it works, amidst the digital-learning hype.

When the kids in Montour school district (PA, USA) turned up to school that day in the fall of 2018, they were in for a surprise. They were told they would begin a brand-new course on Artificial Intelligence (AI). What on earth was AI? And what could it mean to kids in classes 5 and 6?

But this was a serious matter. MIT Media Lab and Media, Arts and Science Department at MIT, had come together and proposed to ‘catch them young’. The idea was to make an early introduction to concepts and practical AI lessons for middle school kids. All students from classes 5 to 8 would go through the AI Ethics program to identify use cases of gender / racial biases, privacy, and fairness. By the end of the 3-day course, they would know if such biases were embedded into the programs they would work on.

Welcome to generation AI. This makes millennium kids look antiquated. This new breed is sensitized to the good of AI and is aware of where it could go wrong. 

That is not all. Montour School district STEM teacher has co-developed a six-week program with Carnegie Mellon Dept of Computer Science called AI in Autonomous Robotics for 7 and 8-grade students. The implementation rigor here is quality stuff as kids are asked to solve real-world problems. 

Amper Music, the world’s first AI music composer and producer, has worked with music faculty at the school to develop a 10-day AI Music program for class 7 and 8 students. This school district is certainly leading the AI drive firing on all cylinders.

A host of universities, AI software firms, educators, and AI experts are coming together like never before to create early engagement for school kids into the AI world. And unlike what most of us would have thought: It is not only about STEM. In fact, the philosophy is to move from STEM to STEAM (with a liberal dose of Art – music, media, entertainment) thrown in for good measure. And this is happening in several pockets across the US.

AI in education sector – AI is here to stay, and the US campuses are already doing it

Across the United States, AI penetration within the education sector is tangible but may not be visible to the untrained eye. While varying in level of experimentation, schools and higher education institutes have embraced the tech and decided to learn how to harness its powers. 

Pittsburg-based Carnegie Learning1 offers AI-based personalized math, applied sciences, and language programs for post-high school students to rediscover learning. The entire program is personalized and self-paced, giving a new approach to STEM learners post-K12 schooling. The results demonstrated in some school districts in Washington and Texas prove the program creates a positive impact.

Duolingo2 is an amazingly popular AI-based customized language learning tool that allows anyone to learn a language. This is based on machine-driven instructions optimized for students based on millions of similar learning sessions held earlier. And most of the learning is for free.

California-based Content Technologies3 is a pioneer in AI and has developed several advanced AI systems for education. The Cram 101 is an AI tool that converts any textbook fed to it into chapter-wise byte-sized summaries, true or false type questions, learning concepts in record time. The company has developed similar tools for different disciplines such as nursing education, high school, and so on.

Some of the interesting outcomes of the approach of starting them young came from a US scientist, Ms. Druga, who built Cognimates, an AI platform for building games and programming robots and training AI models. Cognimates was incubated in MIT Media Labs. 

In a three-year study, where kids were taught to program bots to play games such as Rock and Scissors and build gaming applications using AI. One of the most profound observations came from Druga: When the kids came out after a session and said – “the computer is smart, but I am smarter”.

This was a powerful endorsement of how a young student comes away with a high level of confidence in the programmability of the computer to do what she wants it to do. This clearly establishes the argument about why AI perhaps should be started early on in school.

Next steps in playing this right – How can AI be used in education?

In general, schools and Universities must do the following to stay abreast of the AI curve and help imbue its benefits within the communities.

1.   Create a qualified AI resource team within the institution so they can track AI developments in peer institutes, vendor implementations and research the use cases.

2. Understand own deployments, migration of data systems into the AI realm, define implementation road map and create necessary stakeholder education of the new systems that will come.

3.   Educational institutions should also work with boards, government agencies, and accreditation bodies to define a structured AI curriculum for higher courses. This may require an industry interface also. This combination will create a Special Interest Group -university-industry – regulator group that will work together in ensuring the best interests of all concerned.

4. Faculty training, student and parent education, and awareness programs in terms of how the implementation could affect them need to be made available. Privacy and security rights of all stakeholders are paramount and need to be protected. How the schools intend to ensure data protection as machines become more powerful and open to sharing, receiving data from remote tutors, servers dynamically need to be shared transparently.

The Association for the Advancement of Artificial Intelligence (AAAI) and the Computer Science Teachers Association (CSTA) launched the AI for K-12 Working Group (AI4K12) to define for artificial intelligence what students should know and be able to do.

There are several such movements developing effective programs to deploy at various levels. These can help institutes understand better where AI is headed and how to ride this new technology wave to harness its full benefits.

Start your AI journey with Trigent

AI could well be the elephant in the classroom but if it’s a friendly elephant that can help enrich your life, you wouldn’t complain, would you?

At Trigent, we provide intuitive and easy to use AI solutions that ensuring seamless adoption of the latest technology. With the AI-powered tools from Trigent, you will be able to accelerate your digital transformation initiative in your organization successfully.

Want to know more? Get in touch with us for a quick consultation.

 References

  1. https://www.carnegielearning.com/why-cl/success-stories/
  2. https://www.duolingo.com/info
  3. http://contenttechnologiesinc.com/

Digital Asset Management System – A must-have for all businesses

What are digital assets?

Wikipedia definition: “​​A digital asset is anything that exists in a digital format and comes with the right to use”. For example – video, music, documents, images, presentations, digital tokens (including crypto), data, or anything an organization or individual owns or has the right to use.

As we move ahead with digital transformation more and more businesses are increasingly dependent on digital assets. Today, even existing physical assets such as documents and prints are actively digitized. Digital assets are convenient as they occupy less physical space, are easy to retrieve, and can be transported/transferred easily.

Businesses who have already made the shift to digital assets include, 

  • Legal / Law firms
  • Advertising Agencies, Media houses
  • Broadcasting
  • HR and Recruitment firms
  • Movie Production houses
  • OTTs

Major industries such as retail, manufacturing, import-export houses, insurance, finance, and logistics companies are all in various stages of digital transformation.

With increasing convenience comes its own set of problems, and in this case, it is the management of the digital assets that we create and convert from the existing ones. This is especially true for Business Service companies that create, use and distribute different types of documents and related content. 

How it starts

Every individual and organization starts by organizing their files and their assets in a traditional hierarchical system on their local computers,  USB storage devices,  and of late on the cloud ( Google Drive, email, Dropbox, etc.). Once there is a need to share these assets and use them in collaboration, they resort to shared drives and transfer these assets via email, etc. 

While this kind of organization works on a small scale, the system gets easily overwhelmed with an increase in the number of users and assets.  

Eventually, the challenges present themselves:

  • Single paradigm of classifying our assets – different users / functional-units classify assets differently. E.g. Sales dept will want contracts classified by customers or geography while the accounts teams may want them classified by chronology, billing, risk etc. In short, one size does not fit all.
  • Sharing assets with others – Providing access to “other teams” or third parties is initially simple and can be monitored. However over time, as the content and the teams involved increases, it can spiral into a complete chaos. The most ideal use case would be to provide access to specific assets and probably for a finite amount of time. This brings us to the next point.
  • Security of assets – In 2015, all the first four episodes of the Game of Thrones season surfaced online before it even got aired because the Media outlets provided the episodes for viewing as a part of the review process. This was catastrophic. Sensitive content especially of monetary value needs to be secured and there should be an audit trail to trace any leaks.
  • Version control – While presentation.ppt,  presentation1.ppt, presentation-ver2.ppt would work for an individual or at a small team level, it would require additional tracking effort or worse cause confusion under unwanted circumstances.
  • Automation – Digital assets typically go through a standard workflow including (not limited to) publishing onto websites, pushing to 3rd parties, Watermarking, QA  QC, Approvals etc which could be potentially automated to provide better efficiency.

Enforcement is a key challenge in a discipline-based system and things get cumbersome. There are several Sophisticated DAMs available in the market and when the time comes it is best to get one in place. 

When is the right time to consider a DAM?

Adopting the right technology at the right time is significant for the growth of any business. Here are some points that will help you identify if it is the right time to adopt a DAM in your business 

  1. Are digital assets a significant part of your business?
  2. Does your workforce spend a lot of time looking for files?
  3. Have you had to do a work from scratch when it could have been repurposed from an existing asset?
  4. Are you making duplicate purchases of assets because existing assets cannot be found?
  5. Are unapproved files being used fairly regularly?
  6. Are you losing time validating the “Final version” against the other versions?
  7. Are you spending a significant amount of  time on tasks that can be automated such as watermarking, resizing, transcoding etc?
  8. Does sharing large files require a process which is not as easy as sending email?
  9. Are you finding difficulty in identifying a secure store for your assets?
If you have 3 or fewer “yes”You still have some time. Keep a sharp lookout for the most common cases mentioned. 
If you have 4 – 6 “yes”It is time to start looking for a DAM. It is also a good time to get familiar with a Digital Asset management system. 
If you have more than 6 “yes”Now might be a good time to get your DAM in place.

The losses and risks associated with the loss of Digital Asset Management systems are becoming a standard around the world. The cost of loss and efficiency is real and it has a direct impact on your business.

Hence ensure to be proactive rather than reactive. Also keep in mind that once you have identified the DAM and Vendor, there is still time left (you are the best judge of this) for Deployment, Migration, and User-acceptance. Ensure you plan it well to make this initiative successful. 

Find the right DAM

Once the decision is made to go in for a Digital Asset Management system, there are several choices that need to be made. Broadly they are based on capability/features and cost model.

Features and capability

Consider the following features:

  • Types of assets you will store on the DAM. E.g. Audio, documents, images etc.
  • Attributes of indexing for search and retrieval. E.g. content keywords, Approval status, date, value, vendor etc
  • AI based DAMs can automatically tag features for indexing such as contents of scanned documents, image contents, video and audio content keywords which makes content ingestion a much simpler step 
  • Any automated processes you would like to run on the assets – watermarking, transcoding, resizing
  • Federated Authentication – Consider a DAM that will be able to integrate with your existing authentication system so that the existing system Admin processes will take care of your access management and the users will not have to remember another set of credentials
  • Sharing and permissions – the access various users have to the assets or groups of assets
  • Compatibility with your existing platform and software
  • Any APIs that need to be integrated with the DAM

Buy vs Hire

There are many solutions that can be bought off the shelf, configured, and deployed onto the cloud of local infrastructure based on your requirement. If you already have IT infrastructure and personnel then this is probably a good approach. 

OR

Several DAM solution companies offer a SaaS model where you can just pay a monthly fee and everything is handled. This is typically a good option if you don’t want the upfront expenses or don’t have a dedicated infrastructure team.

Migrate to a Digital Asset Management System

By now you should have zeroed in on the Digital Asset Management system if not already purchased one or subscribed to one.

  • Make sure all the use-cases of all the teams involved are handled. All integrations are in place and all the automated processes are working with their respective types of assets.
  • Ensure you have a buy-in from all the stakeholders involved about the move and set a date.
  • Create the required structure and the attribute lists.
  • Ensure all potential users get their credentials on the new system 
  • Provide training to all the personnelle who will access the DAM
  • Move / Import all existing Assets to the DAM and ensure all new assets are added to the new system.
  • Decommission the old system. This is a very important step as “old habits die hard” and familiarity makes users go back to the older system.

Some popular DAMs

Here are some popular DAMs as per industry leadership sites. Most of these are SAAS-based models. These are pay-as-you-go models and can be a good starting point.

  • Bynder 
  • Canto
  • Digizuite
  • Image Relay
  • Northplains
  • Widen Collective

For the more adventurous ones who already have IT infrastructure and a team that can manage the system, here are some open source options:

  • Islandora 
  • Phraseanet
  • Pimcore
  • Daminion Standalone Basic – The basic standalone is free. They also have a managed service which is a paid model.

A good approach here is to involve your technical team to check on technical skills compatibility and also evaluate the features and their maturity. Even better is to deploy a working copy and test out all the use cases required by all the teams. Most of the open-source projects come with APIs and defined frameworks to extend their functionality.

Confused? 

Get in touch with us for a quick free assessment of your requirement and suggestion for a suitable solution.