Product Review of Most Likely Machine
Most Likely Machine (MLM) is a free website that teaches algorithmic literacy, or how algorithms function and shape our digital world. The site has a classroom yearbook theme and features a series of explanatory pages followed by interactive activities that model algorithmic thinking. In the activities, learners predict which of a set of historical figures (the classmates) will win in three different yearbook categories, such as “Most Likely to Go to a Top University.” Learners are then asked to select a group of traits, assign them to a category, and rank them in terms of importance to that category. Once they’ve gotten all that done, the algorithm runs, and chooses the winners of each category. Learners can then compare their predictions to the algorithm’s results.
Overall, MLM does a good job of showing how various algorithms can be biased and not necessarily give accurate results. However, the selection of terms is a little clunky (you have to choose them one at a time and assign them to each category, but only nine per category). Since you can’t see all the choices in advance, it’s hard to know what should go where. As a result, some selections could end up being not the ones you want. The bio descriptions of the various “classmates” also stray more into opinion than fact, although a lot of this adds to the experience and the lesson.
Most Likely Machine is designed for independent work, but there are lots of great prompts to generate classroom discussion (which learners should be excited to have). Educators should be aware, however, that there isn’t much support for classroom implementation, so be prepared to generate discussion questions and point learners to additional materials should learners want to explore more.
Educators can allow learners to work through the entire site independently and then discuss/debrief with the whole class after they try a few different solutions. Projecting the site and working through it as a class would also be possible. This whole-class model adds great debate to the experience and can unlock a whole new set of discussions and discoveries, from digital citizenship to ethics to SEL. Whatever route you choose, if learners get hooked, they can then return to the site and identify topics they’d like to investigate further. Older learners, in particular, could investigate the algorithmic practices and policies of their favorite sites and design social media platforms or search engines that they feel better respect for people’s privacy.
As an introduction, Most Likely Machine (MLM) does a great job developing learners’ curiosity about, and awareness of, algorithms in a short, focused, and visually appealing experience. The topic will feel instantly relevant to social media-savvy learners and generate a lot of discussion and vectors for further study across content areas. It can get computer science and math learners to reflect on how procedural computational decisions influence behavior, knowledge, and culture. Learners in the humanities get a glimpse of computational processes and will enjoy pondering how algorithms impact fairness and equity. From a digital citizenship perspective, learners will also be encouraged to weigh the benefits and costs of media platforms that depend on these algorithms.
On that note, however, the site is a bit one-sided. The experience doesn’t adequately explain why some might see these algorithms as beneficial, or how design might be able to better account for biases or avoid pitfalls. It also doesn’t address the business models that drive demand for these algorithms. This is left up to the educator, as is any extension or classroom implementation. Adding this context as well as better support materials would round out MLM well.
Overall User Consensus About the App
This site has top-notch design and relevant, unique content. It gets learners making predictions just like an algorithm and then seeing how those predictions play out. The use of historical figures is clever.
Curriculum and Instruction
This experience effectively gives learners the sense that something isn’t right and that the system is biased — and that’s the point.
The site is cleanly designed but doesn’t have any additional resources to extend or deepen learning. It also could be more accessibly designed.