As global debate intensifies around responsible artificial intelligence, product leader and researcher Anmol Aggarwal is examining a practical challenge: how fairness can be integrated into real-world pricing algorithms used by millions of consumers.
Aggarwal, a Product Manager who has built AI-driven pricing and personalization systems at companies including Uber and Adobe, has spent several years researching how algorithmic pricing systems can become more equitable without significantly reducing profitability.
He holds a Master’s degree in Artificial Intelligence and Machine Learning from University of California, San Diego and an MBA from the Haas School of Business at University of California, Berkeley, where he received the Student Always Award for leadership. Aggarwal is also a Senior Member of the Institute of Electrical and Electronics Engineers. Originally from New Delhi, he now works in San Francisco building AI-powered products used by global audiences.
Addressing the Fairness Question in AI
Aggarwal’s work has gained relevance amid increasing international discussions about ethical AI. At the recent India AI Impact Summit in New Delhi, Indian Prime Minister Narendra Modi said AI should be “a multiplier, not a monopoly,” while UN Secretary-General António Guterres called for technology designed with “dignity as the default setting.”
While such forums set broad goals, Aggarwal focuses on operational implementation—particularly pricing algorithms used in digital subscriptions, transportation platforms, and online marketplaces.
Inside the “Black Box” of Pricing Algorithms
Modern digital platforms rely heavily on machine learning models that analyze user behavior, purchasing history, and price sensitivity to predict willingness to pay.
According to Aggarwal, the systems themselves are not intentionally biased.
“The models are simply optimizing for short-term revenue,” he explains. “But when revenue maximization becomes the only objective, fairness failures tend to emerge.”
One common example is the “loyalty penalty,” where long-time customers may pay more than new users because algorithms predict they are more likely to continue purchasing.
“The user who has trusted the platform the longest can end up paying the most,” Aggarwal says.
Research Findings: Fairness With Minimal Revenue Impact
In 2025, Aggarwal published three sole-authored research papers examining fairness in algorithmic pricing. His IEEE paper Fairness-Aware Personalized Pricing evaluated multiple fairness definitions including:
Group parity across demographic groups
Individual consistency among similar users
Protection against loyalty penalties
“Envy-free” pricing where users are not disadvantaged relative to others
Simulations showed fairness constraints reduced loyalty disparities by over 30 percent and cut envy violations by about 50 percent while retaining more than 96 percent of total revenue.
“The trade-off between fairness and revenue exists,” Aggarwal says, “but it is often less than one percent.”
Long-Term Impact and Transparency
Aggarwal’s second study, GTSO++, examined the long-term effects of pricing decisions using a causal uplift framework. Tested on a real-world marketing dataset, it improved conversion rates by 1.8 percent and showed that even a single perceived unfair pricing experience can reduce long-term engagement.
In a third project on explainable AI, Aggarwal demonstrated how pricing decisions can be audited using tree-based models and attribution analysis, tested with sales data from Walmart.
“Regulators cannot audit what they cannot see,” he says.
Aggarwal argues that organizations should track fairness indicators—such as loyalty price gaps and envy violations—alongside revenue and conversion metrics.
“India has the talent to lead responsible AI development,” he says. “The next step is turning that capability into real-world implementation.”
