According Department of E-commerce and Digital Economy about issues of AI applications in e-commerce

On July 18, the Vietnam E-commerce and Digital Economy Agency (Ministry of Industry and Trade) reported on ethical issues when applying AI in e-commerce. Accordingly, currently, ethical challenges and concerns related to AI include concerns about data privacy, algorithmic bias, the delicate balance between user customization and user privacy, as well as the importance of a legal framework to ensure the ethical application of AI in e-commerce.

The application of AI in e-commerce relies heavily on the analysis of massive data sets, including user behavior, preferences, and interactions. While this data-driven approach enhances the efficiency of content creation and recommendations, it also raises significant concerns about data privacy. Consumers are increasingly aware of the value and sensitivity of their personal information, which raises concerns about how e-commerce platforms collect, store, and use data.

Indiscriminate collection of user data for AI applications can lead to privacy violations and unauthorized access. Customers may be uncomfortable with the idea of ​​their browsing history, purchasing tendencies, and personal preferences being used to inform algorithms. Striking a balance between providing a personalized experience and respecting user privacy is important to ensure that e-commerce platforms maintain customer trust.

Algorithmic bias, a common challenge in AI systems, has profound implications for consumers’ experience of fairness and impartiality in e-commerce. AI algorithms learn from historical shopping data, and if this data contains bias, the algorithm can unintentionally perpetuate and even exacerbate existing biases. This can lead to discriminatory outcomes that disproportionately impact certain demographic groups.

Read more: What is digital conversion? 4.0 trend: Digital transformation

In e-commerce, algorithmic bias can manifest in misleading product recommendations, price discrepancies, or discriminatory targeting in marketing. For example, biased algorithms can result in certain users being shown higher-priced products or receiving different promotions based on factors such as race, gender, or socioeconomic status.

Addressing algorithmic bias requires a concerted effort from developers and data scientists to ensure that data is diverse, representative, and free of inherent bias. Regular auditing and transparency into algorithmic decision-making processes are essential to effectively identify and address bias.

“E-commerce platforms must implement robust privacy measures, including clear and transparent data collection policies, user consent mechanisms, and anonymization of personally identifiable information where possible. Communicating and agreeing with users about how their data will be used can foster a sense of control over their online experience.

This also needs to be done carefully, if a platform shows too much insight into its users or spams them with constant recommendations, it can lead to a negative user experience. Striking the right balance will ensure that the application of AI enhances user engagement without crossing the line of invading privacy or causing an unpleasant experience," said the Department of E-commerce and Digital Economy.

To address ethical concerns related to the application of AI in e-commerce, the agency believes that regulatory frameworks play an important role. Regulations may relate to data privacy, algorithmic transparency guidelines, and measures to mitigate algorithmic bias. E-commerce platforms must comply with these regulations and adjust their operations to comply with ethical standards. Ethical considerations should be incorporated into the development process and organizations should proactively engage in relevant ethical discussions.

Overall, the application of AI to support e-commerce brings both opportunities and challenges, especially in terms of ethics. Concerns about data privacy, algorithmic bias, the delicate balance between customization and user privacy, and the importance of regulatory frameworks all need to be addressed. By proactively addressing these challenges, e-commerce platforms can build trust with users, promote a fair and unbiased consumer experience, and contribute to the responsible advancement of AI technology in the digital marketplace. Ethical considerations must be at the forefront as e-commerce continues to grow in the era of AI-powered personalization.

Read more: 5 IT Outsourcing trends businesses need to pay special attention to

Be careful when applying AI in e-commerce

More and more e-commerce businesses are using AI to optimize operations. However, this is a “double-edged sword”.

According to data collection website Statista, Southeast Asia 's internet economy is forecast to reach $363 billion by 2025. And as the sector continues to grow, analysts predict, AI technology is helping to make shopping experiences more seamless, changing retail behavior. 

But like most technologies, the use of AI in e-commerce is a double-edged sword, they warn. And in Southeast Asia, where e-commerce is one of the fastest-growing sectors in the region, the impact could be particularly acute.

AI is not a new technology in e-commerce. Many companies have been using it for years, said Simon Wintels, an expert at consulting firm McKinsey & Company. "It's very common in online shopping, although it may not always be obvious to users. AI is trying to analyze who you are and what your purpose is when you come to a website from the moment you first visit," Wintels emphasized.

AI applications will tailor what users see on a website or app, he said. “It will work in the background as you browse through items on e-commerce platforms. If you add some items to your cart, it will run some AI analysis on other things you might want to add,” Wintels said.

Indonesia currently leads the e-commerce market by size. Statista predicts that the archipelago will account for 42% of Southeast Asia’s e-commerce market by 2030 as the middle class and internet usage there grow.

Indonesia’s e-commerce market is forecast to generate around $160 billion in online retail sales by the end of the decade, up from $58 billion in 2022. However, other emerging markets such as Thailand, Malaysia, Vietnam and the Philippines could also see growth in their e-commerce markets. And this coincides with the growth of AI across the industry.

Read more: Top 10 AI applications that attract the world in 2024

Shopee, Tokopedia and Lazada are the three largest e-commerce companies in Southeast Asia by number of users and sales. Shopee had 470 million monthly visits in August last year, Lazada had 96 million visits in the same month and Tokopedia had 68 million. 

All three companies told CNA that they are using AI to meet the growing needs of their users. A Lazada spokesperson told CNA that the company has launched its first e-commerce AI chatbot, LazzieChat, in Southeast Asia in 2023, in English in Singapore, the Philippines and Indonesia, with a Bahasa Indonesia version to follow soon.

Similarly, Shopee also provides small businesses selling on its platform access to a range of AI tools that help cross-border sellers overcome language barriers in new markets and make it easier for them to build an online presence. 

Tokopedia’s head of communications, Aditia Grasio Nelwan, added that the company uses AI to understand user behavior and improve the experience of using its platform. Through its digital customer service center, Tokopedia Care, buyers and sellers can get their questions resolved quickly without the help of a customer service team.

While Nelwan didn’t share specific numbers on how AI has impacted user numbers or revenue, he noted that during the pandemic, AI improved customer response times by 40% and customer engagement increased by 70%.

However, Dr Unaizah said that businesses should not rely on this technology as AI is sometimes said to be capable of “hallucination”, noting patterns that do not exist and producing meaningless results for queries due to lack of data or bias in the distribution of data used during training.

Read more: Tips for using Google Bard AI to search for information effectively

Additionally, creating ads with the help of AI can pose a major threat if the data fed into the AI ​​algorithm for training is inaccurate or biased. It can miss out on the target audience or discriminate against certain groups in the ad, leading to unwanted legal action.

As AI advances in its ability to predict and influence consumer behavior, people may face greater challenges in making independent decisions.

“This trend has the potential to move us towards a future where algorithms, rather than personal values ​​and beliefs, shape our desires and preferences. Therefore, to address ethical concerns, businesses must prioritize transparency and accountability in the use of AI for marketing purposes,” said Dr. Unaizah.

“AI can be a game-changer in the e-commerce industry when used safely and responsibly,” said Yihao Lim, regional advisor for Japan and Asia Pacific at cybersecurity firm Mandiant Threat Intelligence. “However, AI in e-commerce can be misused or mis-deployed to enable malicious actors to carry out malicious campaigns.”