The risky business of adding AI to the education industry
Data and AI curriculum, from pre-K to Gray, has a lot of growing up to do
According to the World Bank, AI literacy — and we’d also add data literacy — should be developed and integrated into curriculum . It's a matter of fostering a highly skilled and digitally literate teaching workforce and investing in a digitally-accessible ecosystem for educational use. This shift for education is stalling due to the widening digital divide. Learners are grouped into three categories: AI-Empowered, AI-Dependent, and AI-Excluded. Nevertheless, many countries are adding AI topics to their curriculum.
Leveling the data and AI learning playing field has become a necessity. The future of work includes incorporating AI into daily work activities. These McKinsey & Co Report statistics are revealing:
61% of workers currently use or plan to use generative AI
34% of workers expect to use generative AI for more than 30% of their tasks within 1 year
Data and AI skills introduced in the classroom should be easily translatable into the workplace. This is why professionals and practitioners across industries need to assess when to turn toward or away from an AI tool. And if used, they need to determine how much to rely on AI’s outputs. A person’s AI dependency falls within a spectrum that spans tech-first and people-first approaches. Tech-first approaches elevate the AI tool as the go-to resource that’s presumed reliable and trustworthy. people-first approaches, which embrace human-in-the-loop frameworks. AI use is focused, targeted and limited. Each classification has its opportunities and obstacles.
AI-generated classification
The AI-generated classification (>90% AI dependent) refers to a person, company or entity being the most reliant on a series of algorithms and statistical modeling techniques. Human involvement with the designing, development and deliverables of contents rests mainly on the front-end in order to provide input to the AI system, tool and/or platform. For example, DALL-E 3 is an open-source text-to-image generator. All a person had to do was enter a phrase and after a few seconds, an image would appear. The sourcing of the original content and copyright infringement concerns quickly followed. Making changes to the AI-generated image isn’t possible so a user can either accept the generated image or revise their phrase. AI generation is powerful but restrictive option.
AI-assisted classification
The AI-assisted classification (50%-90% AI dependent) involves a person, company or entity first using a series of algorithms and statistical modeling techniques to help them complete a task. Human involvement exists at the design, implementation and outcome stages, but the template used to accomplish that task is provided by the AI system, tool or platform. Beautiful.AI, which started in 2018, uses generative AI techniques to produce powerpoint presentations with slide organization and formatting in mind. The ability to customize the AI-assisted tool’s outputs comes at a price. Regardless, being assisted by AI is allowing the AI to do the heavy-lifting when it comes to organization and structure. AI assistance is most valuable when you don’t know where to start in completing a task. Otherwise, you’re abusing your access to AI-assisted tools and don’t actually need them, you simply want to use them.
AI-enhanced classification
The AI-enhanced classification (10%-49% AI dependent) applies to a person, company or entity sparingly leveraging AI systems, tools and platforms in completing complicated tasks. The words ‘sparingly’ and ‘complicated’ are definitely subjective and depends on the functions to be performed. Human involvement dominates the task completion process with AI serving as the supplement resource if needed. It’s akin to a person having a template or outline of what needs to be completed but they have gaps needing to be more fully developed. We’ve all likely experienced this: knowing the route to a specific location but missing the correct turn. We turn on the GPS navigation app to help us find the best alternative route. As a more commercial example, Zapier released for their Natural Language Actions (NLA). Since Zapier already provides the capability of integrating thousands of online services, NLA gives the user suggestions in building their integration workflows. The lower end is likely where many people will falls while companies will lean toward the higher end of this range as public pressure forces them away from AI-generated and AI-assisted uses.
AI-lite classification
The AI-lite classification (< 10% AI dependent) belongs to a person, company or entity that unwittingly use generative AI systems, tools and platforms. Human-in-the-loop frameworks are prioritized over generative AI techniques. AI approaches aren’t sought-after as the first or second option. There’s a caution-to-use concerning AI systems, tools and platforms. AI use requires clear justification and rationale. Being conscious of AI’s impact makes us cautious in using it. We engage with generative AI tools regularly when we use autocorrect or sentence/phrase completions suggestions while writing an email. While AI-lite seems improbable given the current round of AI hype, the community of AI-lite people will increase due to capitalism — AI approaches will become cost-prohibitive — or due to choice.
There’s no one data and AI system, tool or platform that can solve all our algorithmic-based problems and none of our societal ones. This is the time to explore and become skilled at using people-first AI approaches.
Thanks for reading ! If you're finding this newsletter insightful and valuable, then share with friends, coworkers and the rebel tech curious in your network. Subscribe for free to receive new posts and consider supporting DataedX’s work.