Originally published: Marketing Interactive, 21 July 2025
By Martin Filz, CEO Pureprofile
Hyper-personalisation is the leading focus for brands and marketers, but it comes at the risk of privacy issues. Meta has recently been criticised for exploiting Android web browsing logs on Facebook while Google’s backtrack on cookie deprecation has raised concerns it’s putting advertisers’ interest above user privacy. For marketers, AI-driven market research offers huge potential but must walk a fine ethical line, and the organisations that get it right will win consumer trust.
The rise of AI in market research
AI is seeing unprecedented adoption across all spheres of life and work. Nearly 13 million Australians are actively using AI with usage time up 400% year-over-year. Practically every organisation, across all verticals, is implementing some form of AI across their organisations. For marketers, clients now expect immediate access to insights with no delays. They also expect insights to be faster and cheaper.
AI is reshaping research in profound and powerful ways. It’s enabling predictive analytics, real-time sentiment and automated profiling. AI tools can automatically generate surveys and translate them across multiple languages and perform preliminary data processing and pattern recognition. They can even generate “synthetic data” to reflect realistic audience behaviours, preferences or demographics, enabling brands to test how hypothetical products, messages or market shifts might play out with “virtual consumers”.
New powers bring new risks
One of the most significant concerns is the “garbage in, garbage out” principle, which has become more critical in the AI era. Poor foundational data can contaminate AI systems, creating flaws with major implications, particularly in areas such as synthetic data. If original datasets are biased or limited, synthetic data can amplify these flaws. The quality depends heavily on the original data source and provenance.
A major and growing threat to data integrity is the dominance of non-human traffic online. Today, it’s estimated that up to 50% of internet activity comes from bots. While some bots are harmless or even helpful, many are malicious or misleading, generating fake clicks, form submissions, or fraudulent behaviour. This can significantly pollute datasets, leading AI systems to draw the wrong conclusions about human preferences or behaviour. In advertising, it means brands may optimise campaigns for bot traffic, not real people, wasting budget and skewing performance metrics.
This erosion of data quality is closely tied to a broader issue: declining consumer trust. In an AI-driven advertising ecosystem already flooded with scams, deepfakes and impersonations, users are increasingly wary of how their data is being used. This poses a particular challenge for newer or lesser-known brands, which often struggle to establish credibility and differentiate themselves from bad actors.
As a result, data security is becoming a competitive differentiator. Poor practices will damage brand equity, but businesses can increasingly build brand loyalty through transparent data practices. Emerging norms include zero-party data, value exchange models and the rise of “privacy by design”.
Best practice approach for AI insights
Compliance and transparency are becoming resource-intensive but necessary investments. As one CMO reported, marketers are spending as much time on compliance for a campaign as they were saving, from an AI point of view. To ensure effective and ethical AI use in the insights landscape, there are several key practices that successful organisations are implementing.
The most successful emerging approach is tech-driven service integration, which combines advanced AI tools with human expertise. AI models are used in real-time while maintaining human interpretation and strategic guidance.
Another is embracing more multi-faceted skills development. This means moving away from siloed specialists toward professionals who can combine analytics, product marketing, business strategy and storytelling skills. Around 50% of insights professionals are actively reskilling for AI collaboration, focusing on prompt writing rather than just traditional research techniques.
Successful implementation also prioritises data quality and fraud detection (40% of researchers now prioritise this) while using smart automation for survey creation, translation and analysis. The emphasis is on ensuring foundational data quality is maintained even as processes become more automated.
Measuring the actual business impact of AI, rather than just cost savings or speed improvements, is also important, particularly as AI comes under increasing scrutiny. One organisation calculated $3.70 ROI for every $1 invested in GenAI.
Agile experimentation remains vital. Insights organisations need to be continuously piloting new methods and scaling what works, rather than sticking with traditional approaches. Only the agile will survive in this rapidly changing landscape, particularly as new technologies as well as new regulations come into play.
Creating proactive transparency standards
Rather than wait for regulation, marketers should get ahead of the curve and take a more proactive approach to transparency. This means creating guidelines, standards and transparency tools. This will not only lead to greater consumer confidence but also a higher chance of sensible regulation.
Organisations should always consider going beyond basic compliance. Consumer profiling for example is an ethical grey area. Just because you can predict behaviour, should you always act on that? There’s a need for empathy and restraint with data use.
In an environment where it’s easy to create fake brands and misleading content, authentic brands that maintain transparent practices and build genuine trust will have significant competitive advantages in maintaining customer loyalty and brand equity.
