The balance between privacy and innovation

Stefanie Skomorucha
6 min
|
May 25, 2018
On May 25th, Europe introduced some of the toughest online privacy rules in the world. While it will force businesses to take greater care of the data they collect, there are also certain implications for the development of emerging AI and machine learning technologies.

These new regulations are aimed to give internet users more control over what is collected and shared about them, and consequently, will heavily punish companies that violate these regulations. So what does that mean for the machine learning and artificial intelligence? Where will they fall when it comes to GDPR?

Unfortunately, there still seems to be a lot of confusion when it comes to knowing exactly how GDPR will affect both machine learning and modern artificial intelligence methods. With accountability in this era being vague, it is estimated that 46% of companies believe that they would not be held responsible for the data collection across digital properties.

The truth is, now that the GDPR is in effect, businesses will be held responsible for the way they collect and use data, and unpreparedness could potentially result in brand detriment.

So, let’s break this down further.

Will the GDPR make data collection harder?

The Economist called personal data “the world’s most valuable resource” ahead of oil. And as you should be aware by now, the General Data Protection Regulation has gone into effect across the entire European Union, also applying to all personal data that is collected within the EU.

So how can a business that capitalizes on the power of artificial intelligence and data collection ensure that they are GDPR compliant? Well for one, organizational processes must reflect a standard for data governance to ensure that the process of gaining consent are appropriately captured, and that all parties involved have a “right to explanation.”

Under these new regulations, businesses operating within the European Union are required to introduce systematic changes. According to Bryce Goodman, researcher at Oxford Internet Institute, he explains that the EU has been investing heavily in AI research and machine learning technologies, but that these laws will inadvertently impede their development and use of them in Europe.

Below are just two ways that it will impact the use of AI in European-based businesses:

  • Human review is now necessary to explain algorithmic decisions, resulting in making it more difficult for companies to automate certain processes that use AI.
  • Customers will be able to demand that companies erase all copies of their personal data, which will undermine the integrity of companies’ algorithmic models.

In theory, this has presented some real issues for businesses who rely on the complexity of high-quality machine learning, who often keep their algorithms closely guarded as it is their competitive edge.

Will there be a trade-off between transparency and accuracy?

It could be claimed that with these particular aspects, the human review of algorithmic decisions could force companies to use less accurate AI systems, resulting in potentially unfair decisions. Especially in cases where AI is being used to automate processes and provide decisions for a wide variety of functions such as: marketing models, email filtering, voice recognition, content classification, and marketing optimization, just to name a few.

Popular AI powered voice recognition software

In a report created by the Center for Data Innovation, they argue that due to the nature of these kinds of decisions, they should not be based on whether it was made by a human or an algorithm, but how fair and unbiased the resulting decision is. By using AI algorithms that are adjusted over time for unintended biases, which are now becoming prohibited, will that ultimately affect the consumer and business?

While this is one side of the argument, another perspective is that AI-powered projects should take certain precautions. Instead of basing GDPR requirements on how a particular decision was made, the EU should only reinforce requirements for transparency, evidence, oversight, or technology-neutral explanations. By increasing their transparency and trust between the company, and their partners and customers, it would create a better regulatory environment without reducing algorithmic advancements.

Will the EU still lead in AI after the GDPR?

Artificial intelligence and machine learning technologies will come across difficulties in privacy obligations, and these technologies will come at a cost in terms of innovation and productivity.

The ‘Center for Data Innovation’ further argued that EU policymakers need to realize that failing to amend the GDPR for AI and machine learning technologies will reduce its impact on AI developments, and it will surely put Europe at second-tier status in the emerging algorithmic economy.

So, if these stringent regulations are not somewhat revised, what would it mean for the firms in the EU that develop or use AI? Would it put them at a competitive disadvantage to North America and Asia who are both competing for global leadership in AI?   

Be prepared

Regardless of the issues around the wording and practicalities of the GDPR, businesses need to be prepared.

Even if you were to take machine learning, and override automated algorithms - which is already being done in many cases, the best way to protect your company from non-compliance would be by “simplifying” machine learning projects. This means that as the marketer you would be able to easily explain your analytics and how they were generated. By putting a compliance process in place it would also ensure fair and proper use according to the new regulations.

Nevertheless, machine learning and AI are still the future of marketing, the introduction of the GDPR just means that the “free-for-all” is over. Some can see it as detrimental to everything that AI and machine learning could potentially do, but maybe the GDPR is simply introducing a higher level of accountability for humans and the machines that they use to make decisions.

GDPR, AI, Machine Learning, Data, Privacy

Want to know more about our tech

Contact us

Related articles

300 Park Avenue, NY 10022 - New York, USA | Lehargasse 7, 1060 - Vienna, Austria | Rokin 86, 1012 KX - Amsterdam, The Netherlands
© 2018 Dashmote | All rights reserved