Managing Data – Software Services and AI Legal

Managing data is an essential part of the operation of a growth business. It’s a cliché often bandied around that today data is more valuable than oil. But as with oil, it’s only how the resource is used that defines its value. Whereas oil can be relied upon to produce energy in all circumstances, data cannot be relied upon to produce useful insights at all times. Therefore, the means and purpose by which it is processed becomes all the more important. Given its potential, it comes as no surprise that initiatives, public and private, for managing data more effectively are commonplace. The legal sphere attempting to regulate this burst of energy gets more complex by the day. Here is our introduction to some general issues you may face when managing data for profit, or to simply improve the running of your business.

GDPR and Brexit

Before GDPR came into force in all EU member states on 25 May 2018, the ICO commissioner stated in the ICO’s March 2017 paper, Big data, artificial intelligence, machine learning and data protection, that ‘it’s clear that the use of big data has implications for privacy, data protection and the associated rights of individuals… In addition to being transparent, organisations… need to be more accountable for what they do with personal data’.

At the end of the Brexit transition period (January 1^st 2021), the GDPR and parts of the Data Protection Act 2018 became part of a new body of retained EU law. Essentially replicating the old regime in the UK. Data protection legislation in the UK is now comprised of the UK GDPR and the DPA 2018. From a UK perspective the GDPR operating in the EU will be known as the EU GDPR.

As the EU GDPR will continue to have extra-territorial effect (Article 3, EU GDPR) it may continue to apply to UK organisations who act as controllers or processors and have an establishment in the EU, or who offer goods or services to data subjects in the EU; or monitor their behaviour, as far as their behaviour takes place within the EU. UK businesses could therefore find themselves subject to parallel data protection regulatory regimes under both the UK GDPR and the EU GDPR.

Are you managing data as a processor or controller?

If offering a service, for example a software platform that allows companies to process personal data, then it would often be prudent to ensure you are defined as a data processor, and not a data controller, for data protection purposes. This is because, as opposed to data controllers who bear primary responsibility for the personal data involved, data processors have less obligations under data protection laws. Processers are essentially processing data under the instructions of the data controller. Whilst a data controller determines ‘the purposes and means’ of processing the personal data (Article 4(7), UK GDPR). A helpful way of thinking about it is that a data controller has direct duties to data subjects whereas a data processor only has duties to the data controller.

The distinction between controller and processor in an AI context was first considered in the ICO’s July 2017 decision on an agreement between the Royal Free Hospital and Google DeepMind. Under the agreement DeepMind used the UK’s standard publicly available acute kidney injury algorithm to process personal data of 1.6 million patients. The ICO ruled that the hospital had failed to comply with data protection law and was ordered to perform an audit on the system. The hospital’s law firm, Linklaters, concluded in the hospital’s audit report, Audit of the acute kidney injury detection system known as Streams, that DeepMind had been properly characterised as a data processor. This was because Streams ‘does not use complex artificial intelligence or machine learning to determine when a patient is at risk of acute kidney injury. Instead, it uses a simple algorithm mandated by the NHS’. It was therefore the lack of complexity involved in the ‘means’ of processing the personal data which meant that DeepMind were considered to be a data processor. A complex algorithm would have constituted a level of agency on DeepMind’s part which would have rendered their processing that of a data controller. It was deemed, however, that their services were simple enough to be doing nothing more than following the hospital’s instructions. This grey area should be of concern to anyone planning to use AI to analyse data. Make an algorithm too complex and you may take on the liability of a data controller and hence liability towards data subjects.

Anonymisation

Managing data to make it anonymous would fall under UK data protection laws. This is because the purpose with which the personal data was originally collected needs to be aligned with the purpose that it is later anonymised for. There are certain circumstances in which collecting personal data to begin with is not necessary and, if still useful, highly desirable for businesses wishing to process the data as they wish. If the data is originally collected in an anonymous format, then UK GDPR no longer applies. As GDPR states at recital 26, ‘the principles of data protection should… not apply to anonymous information, namely information which does not relate to an identified or identifiable natural person or to personal data rendered anonymous in such a manner that the data subject is not or no longer identifiable’.

In an ICO report, Anonymisation: managing data protection risk code of practice, the ICO lists anonymisation as one of its six key recommendations for AI. It states ‘organisations should carefully consider whether the big data analytics to be undertaken actually requires the processing of personal data. Often, this will not be the case; in such circumstances organisations should use appropriate techniques to anonymise the personal data in the data sets before analysis’.

Profiling and automated decision making

AI’s ability to uncover hidden links in data about individuals and to predict individuals’ preferences can bring it within the GDPR’s regime for profiling and automated decision making. Article 22(1) states that ‘a data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly affects him or her’.

However, this is qualified by article 22(2) which states that this right does not apply to a decision that ‘(a) is necessary for entering into or performance of a contract between data subject and data controller; (b) is authorised by… law to which the controller is subject and which also lays down suitable measures to safeguard the data subject’s rights and freedoms and legitimate interests; or (c) is based on the data subject’s explicit consent’.

This is further qualified: ‘in the cases referred to in points (a) and (c)…, the data controller shall implement suitable measures to safeguard the data subject’s rights and freedoms and legitimate interests, (including) at least the right to obtain human intervention on the part of the controller, to express his or her point of view or contest the decision’. Having automated decision making within software performing data analysis can therefore introduce new obligations. Such obligations often being onerous for a data controller. This can include it being necessary to perform a Data Protection Impact Assessment or getting explicit consent from data subjects.

Other suggested compliance mechanisms

The ICO makes five recommendations for using AI to analyse data:

• Privacy notices.
• Data protection impact assessments – embed a privacy impact assessment framework into data processing activities to help identify privacy risks and assess the necessity and proportionality of a given project.
• Privacy by design – implementing technical and organisational measures to address matters including data security, data minimisation and data segregation.
• Ethical principles.
• Auditable machine learning algorithms.

Treasure trove

Finding new and innovative ways for managing data is a treasure trove many wish to unlock. It is important to be wary of the growing regulatory landscape underpinning the sector. The world was shocked by the accusations made against Cambridge Analytica and making sure you display compliance is a must for maintaining a good reputation and attracting clients. With Brexit comes the potential for the complexity inherent in potentially diverging legal regimes. Being up to date on the development of the Privacy and Electronics Communications Regulations (PECR) will also be useful. Read our blog on PECR here.

EM law specialises in software technology law and data protection law. Get in touch if you need advice on data protection law or if you have any questions on the above.