Skip to content

New AI ‘assistant’ being tested by NYC teachers to create lesson plans

1787 Constitutional Convention Generative prompt on Adobe Firefly

An artificial intelligence program from South America that creates lesson plans for teachers is now being piloted in some Brooklyn high schools, The Post has learned.

Teachers in northern Brooklyn are being trained on how to use YourWai, an AI “teaching assistant” created by developers in Colombia for the company Learning Innovation Catalyst, Superintendent Janice Ross announced at a parent council meeting Wednesday.

“Teachers spend hours creating lesson plans. They should not be doing that anymore,” said Ross, who oversees high schools in Sunset Park, Park Slope, Williamsburg, Bed-Stuy, Brownsville and East New York.

Teachers can enter students’ needs and the standards they want into the app and a lesson plan will be generated, Ross explained at the Citywide Council for High Schools meeting Wednesday. She called it a “game changer” that will give teachers more time to “think creatively” and less time spent creating curricula.

“Will it take time, will we have to adjust it if you want more of this, want more of that? Yes,” Ross added. “We can never replace humans.”

But some educators and observers warned AI lesson planning is concerning on many levels: fledgling AI systems have been shown to lean left and present absurd historical inaccuracies, and lesson planning itself is an essential and creative teaching activity.

“The mistake is to see lesson planning as drudgery that can be off-loaded to a machine,” Brooklyn College and CUNY Grad Center education professor David Bloomfield told The Post.

“Using AI, like long-available commercial or free materials, shouldn’t be a crutch yielding sterile, depersonalized work, but a tool that teachers will need to be taught to use if it’s to be effective,” he added.

Google’s Gemini was recently blasted for creating “diverse” images that were not historically or factually accurate — like female popes and Native American Founding Fathers.

Adobe’s Firefly recently generated depictions of black Nazis.

A study conducted by New Zealand professor David Rozado for the Manhattan Institute last year found “instances of political and demographic bias” and “left-leaning” undertones in ChatGPT responses.

Morning Report delivers the latest news, videos, photos and more.

Please provide a valid email address.

Morning Report delivers the latest news, videos, photos and more.

By clicking above you agree to the Terms of Use
and Privacy Policy.

Never miss a story.

Rozado told The Post this week that most AI algorithms display bias in response to political topics.

“Human led fine-tuning” is needed to align AI with a certain set of values, he said.

“Another potential problem with AIs creating lesson plans for students is the relatively high frequency with which AIs display hallucinations/confabulation,” he said. “That is, they make stuff up that is simply not true.”

Rozado presented the potential scenarios that could play out: AI can reinforce a human’s biases, a human can correct or learn from the technology, or, “in a more disturbing scenario, a biased system could present incorrect content in such a persuasive fashion that it convinces the user of its veracity,” he said.

“We know that AI can sometimes be wrong, and it must be used cautiously,” Queens Councilman Bob Holden, a member of the City Council’s Committee on Technology, told The Post.

Today's News.
For Conservatives.
Every Single Day.

News Opt-in
(Optional) By checking this box you are opting in to receive news notifications from News Rollup. Text HELP for help, STOP to end. Message & data rates may apply. Message frequency varies. Privacy Policy & Terms:
This field is for validation purposes and should be left unchanged.