However, these institutions have long been dubbed laggards when it comes to technology, innovation, and the speed at which they can digitally transform. Much of this is due to the legacy infrastructure in place, the regulatory landscape in which they operate, and security and governance protocols they have been hamstrung by. This means that data is not driving the valuable innovation it can do to improve automation, decision-making and risk management.

In comes synthetic data. This is not ‘real’ data created naturally through real-world events. It is ‘artificial’ data that maintains the same statistical properties as ‘real’ data, generated using algorithms. Whether the aim is to make data available across an organisation or accessible to third-party partners it drives speed to innovation within financial services.

This is already happening as the first banks start to roll out synthetic data across various use cases, from testing to AI model training to cloud migration projects. But in 2023, I believe the sector will open its eyes to the notion of synthetic data and how it can fuel growth, support overcoming longstanding obstacles, and totally rejuvenate the way financial services institutions meet and exceed the ever-evolving requirements of customers and regulators.

Revolutionising data privacy 

According to Gartner, synthetic data will enable organisations to avoid 70% of privacy violation sanctions. Financial data, such as consumer transaction records, account payments, or trading data, is sensitive personal data subject to data protection obligations and is often commercially sensitive.

Structured synthetic data has the potential to revolutionise the way financial institutions use data securely. Because this data preserves the statistical properties of real-life data, the strict privacy and security protocols that have previously blocked innovation can now be navigated with synthetic data. So, because no real individuals can be identified from the synthetic data, data protection obligations, such as GDPR, do not apply. This will undoubtedly be top of mind in 2023 for business leaders, with the fifth anniversary of GDPR in May.

Since privacy compliance and information security regulations will no longer be an issue, the new artificially generated data can then be used to create new revenue streams. The banking sector can take their Open Data and data monetisation strategy even further in 2023 since synthetic data will enable them to package this data and sell it to third parties without the need for express consent.

Seamless cloud migration

There’s no doubt that the organisations that are succeeding in these trying times are those that can rapidly scale via the hybrid or public cloud. But well-regulated industries like banking and financial services have been reluctant to go all-in with the cloud. I get it. As soon as data leaves the company campus and servers, the control is lost. Synthetic data allows for a rapid, cross-organisational migration to the cloud without any of the added risks. Something that financial organisations can use to great effect in 2023.

Instead of pseudo-anonymised data (created by traditional processes such as masking and anonymisation) that can still lead to re-identification or redacted data that loses most of its utility, with synthetic data generation, the dataset is totally new and holds no ties to the original. If used, in 2023, financial services can train on their real datasets on-premise – even behind the walls of separate departmental silos. Then, the artificial data can be released into the cloud. And since there’s no personal information in it, the synthetic data can be shared across silos within the organisation — allowing for cross-organisational strategy, insights and analytics like never before.

The commercial impact of generative AI

Generative AI underwent a huge step change in the latter half of 2022. Teams from OpenAI through to StabilityAI have been creating models that can conjure hyper-realistic text and images from seemingly thin air with very minimal verbal prompts. The realism of the responses you can get from these models is in some cases quite creepy and like nothing we’ve seen prior to this year.

So how will this development impact business and society? The jury is still out, but what we do know is that these teams are making these models available for anyone to play with for free right now creating the perfect test ecosystem for developers, hackers and anyone who’s curious to test their ideas.

I am certain that in 2023 we will start to see businesses forming around these tools. For example, there are already examples of text or formula auto-completion tools being embedded into Microsoft Office software that could greatly improve productivity and speed up learning curves of users. These types of efficiency-improving tools have the potential to impact businesses much further afield than just financial services.

There are certainly some concerns and legal challenges that still need to be overcome before this technology can be commercialised. Who owns the output of one of a code auto-completion model if it was trained on data under different licences? Who owns the copyright to images generated from a model?

Despite these challenges, there is huge potential in this technology, and I believe we will all be hearing much more about it in 2023.