Suggestions
Gary Liang
AI Startup Builder and Consultant, Former Management Consultant
Gary Liang is the Founder, CEO, and CTO of Bloom AI, an educational technology startup focused on providing AI tutoring solutions for universities and high schools. In addition to developing AI-driven educational tools, Bloom AI also offers consulting services to educational institutions and startups regarding strategy and artificial intelligence.
Liang has a robust background in management consulting, having worked at McKinsey & Company, where he advised clients across various sectors including telecommunications, banking, and technology. He has also held roles as a business analyst at SolveXia, a no-code automation startup, and has experience in the financial sector with firms like Goldman Sachs and IMC Financial Markets.
His academic qualifications include a first-class honors degree in economics and pure mathematics, with a specialization in random graph theory. Liang has also founded a high school tuition center, demonstrating his commitment to educational initiatives.1
Highlights
Amazing to hear Sal Khan, Jordana Hunter, Bryan Lee and Rose Luckin speak about the future of education in an AI-powered world at EduTECH Australia.
Can generative AI harm learning?
The answer is yes, if implemented blindly.
A brilliant paper was recently published by Hamsa Bastani, Osbert Bastani, Alp Süngü, Haosen Ge, Ph.D., Özge Kabakcı and Rei Mariman, researchers from the University of Pennsylvania and Budapest British International School. Reading this study gave me even more conviction of the importance of what we're doing at Bloom AI.
They undertook a randomised controlled trial on a group of high school mathematics students in Turkey. The students were split into 3 groups:
- Control: No AI, just course textbooks and notes.
- GPT Base: Access to GPT-4 chatbot similar to ChatGPT.
- GPT Tutor: Access to a carefully crafted GPT-4 chatbot, with a prompt which contained detailed solutions and instructions to not reveal answers directly.
Students' performance in the assisted practice problems substantially improved in the GPT Base group (by 48%) and in the GPT Tutor group (by 127%). The big result was that unassisted exam performance actually degraded by 17% for the GPT Base group. The negative effects were largely mitigated by the safeguards in the GPT Tutor group.
My key takeaways:
- Blindly implementing off-the-shelf AI tools may harm learning
- We need AI specialised for education
- Give students guidance on how to use generative AI
- AI is like the calculator… but it's also not
I also suggest some further avenues of exploration.
Check out my write-up here: https://lnkd.in/gGM78ZFk
