Connect with us

Business

BANKING ON BOTS – MITIGATING ALGORITHMIC BIAS IN THE FINANCIAL SERVICES

Published

on

by Clare Joy, Strategy & Expansion Lead at Onfido

 

When developing new technologies, we must ensure that they operate fairly. At a time when identity is increasingly being used as the key to digital access, any technology based on identity must function fairly and equally for everyone, regardless of race, age, gender, or other characteristics leading to human physical diversity. While digital services have proliferated across many industries, this issue is particularly relevant in the financial sector, as Covid-19 accelerates a shift towards automated platforms delivered remotely by banks and other providers – with biases in AI having stark implications for unfairly rewarding certain groups over others.

 

How does AI bias creep into machine learning models?

Algorithmic decision making relies on machine learning techniques that recognise patterns from historical data. While often successful, it can pose a significant threat when these patterns are based on biases found in the data – these can emerge in two scenarios. First, a standard machine learning model can incorporate the biases found in the data during training. This can lead to subsequent predictions being made based on these biases.

Clare Joy

The other is that although the data is not necessarily biased, there could simply be less data available from a minority group for training. When there is less data to work with, especially with modern machine learning techniques, this is more likely to lead to modelling inaccuracies.

When this happens, it can have a real world impact. For instance, a criminal justice system in Florida has been found guilty of mislabelling African-American defendants as ‘high risk’ at a much higher rate than white defendants. We also saw Amazon discontinue use of an AI-powered recruitment platform which was shown to prioritise male applicants based on the language they used in their CVs.

In the financial industry, many processes that underpin much of society – from credit assignments to mortgage approvals – are simply not as fair as they should be because decisions are based on historical biases. Every individual or group should have the same set of opportunities, regardless of gender, age and ethnicity. For those of us that work with machine learning models, it is imperative that we try to minimise cases of unjust bias and understand how bias arises in our models.

 

Measuring and mitigating bias

Several tech giants are already attempting to do this by releasing supporting software for various parts of the machine learning lifecycle to mitigate biases. For instance, Google released multiple fairness diagnosis tools and a library enabling the training of fair models. Microsoft and IBM have also released tools for assessing and improving algorithmic fairness. However, it is incumbent on all businesses to optimise their own AI processes to eliminate bias.

This is something that we at Onfido focused on in the Information Commissioner’s Office (ICO) Privacy Sandbox, where we systematically measured and mitigated algorithmic bias in our artificial intelligence technology, with a particular focus on racial and other data related bias effects in biometric facial recognition technology. This closed the difference in performance between ethnicity groups for our facial recognition algorithm, which included achieving a 60x false acceptance improvement for users in the “Africa” category.

Part of the solution means companies using AI must review their machine learning models to ensure that they are not using biased data. Regularisation during training is one way of adding fairness, although this assumes that the model and relevant data are both available for a particular vendor or practitioner. By mathematically denoting a notion of fairness, it is possible to optimise for the chosen fairness constraints by adding them to the objective function.

Alternatively, pre-processing the training data sets means that features and sensitive information are decorrelated before training, while having a minimum impact on the data or decision rules. This is particularly applicable when there is no access to both the data and the training pipeline. Another strategy to obtain fairness is post-processing which is done by adjusting the classifier after training, when the pipeline is either unavailable or re-training is costly. By recalibrating the classifier after training, the threshold is set so that it maximises a certain fairness criterion.

 

Championing fairness

Ultimately, by formalising a mathematical notion of algorithmic fairness, we give ourselves a way to remove biases at the data stage, during the model training stage and through post-processing adjustments. Integrating and monitoring fairness constraints in this way can ensure that algorithms provide the same level of opportunity for every group throughout society.

Deploying machine learning models is a responsibility as well as a tool for all businesses that work with AI. While removing bias is essential for improving customer onboarding and user journeys, we have an ethical imperative to minimise cases of AI bias and understand how it arises in our models. In particular, global financial services with such influence over wealth distribution must ensure they do not exhibit biases that could hinder the opportunity of certain groups, which is critical as we enter a privacy-preserving world.

 

Business

Ransomware chokes COBRA: How AI-powered data analysis can support financial services’ plight

Published

on

By Toby Butler, Financial Crime Solutions Manager at Ripjar

 

Ransomware attacks are on the increase in the United Kingdom. Most of the British Government’s COBRA meetings have been convened in response to ransomware attacks, showing how cybersecurity breaches are as pressing as national emergencies and crises. The National Cyber Security Centre’s (NCSC) annual review found this year that the country was hit by 17 ransomware incidents that were so impactful they “require a nationally coordinated response”. That extends to the financial services sector, which saw an increase of ransomware attacks with 55% of organisations hit in 2021.

Where does this leave the sector and how can artificial intelligence and machine learning be instrumental in understanding the risks companies face against future ransomware attacks?

Toby Butler

Company information is being stolen and sold to different threat groups, who prey on the individuals in that organisation who are more likely to pay them. The UK is one of the most cyber-attacked countries in the world and the Government has been criticised for being “ill-equipped” to deal with this exponential rise of fraud cases.

 

Ransomware-as-a-Service

Ransomware is one of the most common forms of cybercrime. Fighting it has become one of the biggest problems that organisations today face during their everyday operations. For instance, Malware (malicious software) encrypts the files of a single computer, then works its way through an entire network to reach the server and inflict maximum damage. Company information is being stolen and sold to different threat groups, who prey on the individuals in that organisation who are more likely to pay them.

When these attacks occur the victims, more often businesses, are left with minimal options. If they have substantial backup solutions already in place, they can attempt to restore the encrypted data to their servers. But if that data isn’t already secured elsewhere, they may need to pay a ransom to the criminals behind the attack. Thereby allowing the business to function once again and restoring their reputation. The cost of paying the ransom will feel considerably smaller compared to starting a business again from scratch. Sophos’ State of Ransomware in Financial Services 2022 report found that 52% of financial services organisations paid the ransom to restore their data, the average remediation cost in financial services was US$1.59M.

Cybersecurity Ventures estimates that ransomware is set to cost global businesses more than $256 billion by the end of 2031. By that token, organisations need to be extremely mindful of the potential threats they may face. Businesses need to understand the methodologies these hackers use, to address the weaknesses within their domain and take measures to isolate and prevent further ransomware attacks from happening again.

 

The rise of WAMs

According to a recent report by security firm CyberSixgill, 19% of the 3,612 cyberattacks that took place in 2021 were traced back to Wholesale Access Markets – or WAMs for short. WAMs are, in essence, underground internet flea markets. These markets are where aspiring attackers come to purchase network access from threat actors – the individual or entity involved in carrying out the cyber-attack. Types of threat actors include insiders, cybercriminals, rival organisations, or even nation states stealing data.

WAMs sell access to multiple compromised endpoints (or pathways) for around 10-20 dollars. Researchers found that WAMs listed access to approximately 4.3 million compromised endpoints in 2021, which include access to both provider and enterprise software (for example, an organisation’s Slack channel) up to 180 days before the attack itself took place. This shows how long these compromised endpoints remain undetected without proper internal analysis.

 

How can Financial Services stay ahead of the curve?

The use of Artificial Intelligence (AI) and machine learning is undisputed across modern businesses and sectors, and continues to revolutionise processes across the board. AI is a significant player in the financial services industry, building the ‘cyber-wall’ against nefarious users. It gives organisations optimal insights into reducing the likelihood of a ransomware attack in the future.

Namely, AI and machine learning collects and analyses vast amounts of messy (structured and unstructured) data from disparate sources. The challenge for the sector is to understand the volume and variety of the raw data collected from any source to build better protection in the future.

Structured information could be best understood as the clear data we see in a table. For example, the following attendees made a business meeting: first name – Joan, surname – Smith, age – 46. But unstructured information is information presented in a complex manner. For example, ‘there were five people who attended the business meeting, one of whom was forty-six and called Joan Smith’. Naturally, due to the complex nature of the prose, it would be more difficult for a machine to process that data into a digestible format for further risk analysis. This is where AI continues to prove invaluable.

AI uses natural language processing to understand the information provided on the web. As the software continues to evolve, natural language processing reads the information in a way a human would to extract the key information from the text. By incorporating AI and machine learning within an organisation’s IT infrastructure, companies operating within financial services can be better equipped to handle cybercrime.

These tools are flexible and adaptable, they can be configured to analyse different types of data from different sources to curate key insights. This collated information provides a better analysis of the organisation’s exposure, allowing them the opportunity to get upstream in preventing future attacks. This kind of approach is essential to processing listings on WAMs.

The power to analyse data to identify weakness is vital in the battle against cybercrime. It gives organisations a better understanding into what they could expect to see in the future. Hosting the correct data, and with the analytical skills, financial organisations can gain a better understanding of the methodologies and weaknesses in-house that attackers use and exploit to hold them to ransom. Organisations can then use this as a reference to pinpoint compromised endpoints, giving them a chance to reduce access before this route can be exploited and ruin their business.

With cybercrime and ransomware continuing to remain prevalent, it’s vital that financial services companies understand how they can get ahead of the curve and build a robust security platform within their IT infrastructure that can withstand an attack. In 2022, a ransomware attack occurred every 40 seconds. The mindset for the sector needs to be one of when, not if.

Organisations need to be thinking about an attack now – before it’s happened. Pre-planning and preparing for the worst possible outcome from future threats and adversaries. The introduction of AI and machine learning in the fight against cybercrime is a must, and the sooner the industry gets behind in implementing AI, the safer it will be through the next decade.

 

 

Continue Reading

Business

SVEA BANK ACQUIRES AREX’S FINTECH OPERATION IN FINLAND

Published

on

By

AREX Markets, the data-driven FinTech company that drives financing costs down for SMEs and enables them to get paid quicker, has announced the sale of its Finland operations to Swedish payment and financing institution Svea Bank.

With the deal, Svea will further strengthen its position as a corporate financier, as AREX’s c.1200 Finnish customers and partnerships in the areas of financial management and financial management software will be transferred to the bank’s portfolio. The Finnish operation of AREX has financed over EUR 500M worth of invoices.

AREX’s Spanish and UK operations remain unaffected and remain focused on building embeddable financing products for third party platforms. Customers in Finland have been informed of their transition, and their contracts and service details will port across to Svea.

Svea is reshaping the playing field of corporate finance in Finland, and taking on the operations of AREX in the region is a natural step to strengthen their own business and at the same time offer AREX’s partners and customers an easy path to a wider range of services than before.

“Over the years, Svea has grown a lot also through business transactions, therefore acquiring AREX’s business operations in Finland was a good and natural solution for us. In addition, the deal is pleasant for us at Svea because the focus of our activities is to help partners and customers succeed – offering AREX’s partners and customers a wider range of services is exactly that,” says Pasi Väre, country manager of Svea in Finland.

The deal also brings new opportunities for AREX to focus on the UK and Europe in its roll out of embeddable financing products, which can be white-labelled by neobanks, ERPs and accounting software alike. The business is seeking to bridge the liquidity gap faced by most small businesses in the face of a recessive economic climate.

UK SME’s can continue to access AREX’s core invoice financing product through the Xero marketplace.

“For us at AREX, this is a great step: we are developing a stronger presence in the field of embedded finance, which is underpinned by our sophisticated marketplace software, our strongest point,” says AREX’s CEO, Airto Vienola.

“For the AREX team it was extremely important that we find the best possible corporate financier to take care of the business’ customers and partnerships in Finland. Svea convinced us with their customer and partner-centric approach”, adds AREX’s co-founder Perttu Jalkanen.

Continue Reading

Magazine

Trending

Business14 hours ago

Ransomware chokes COBRA: How AI-powered data analysis can support financial services’ plight

By Toby Butler, Financial Crime Solutions Manager at Ripjar   Ransomware attacks are on the increase in the United Kingdom....

Banking21 hours ago

How Banks Can Boost App Innovation, Speed and Compliance

Steve Barrett, Senior Vice President of International Operations, Delphix  As new finance and banking applications disrupt the market each day,...

Business21 hours ago

SVEA BANK ACQUIRES AREX’S FINTECH OPERATION IN FINLAND

AREX Markets, the data-driven FinTech company that drives financing costs down for SMEs and enables them to get paid quicker, has...

News21 hours ago

ICICI Lombard and AU Small Finance Bank announce Bancassurance tie-up

ICICI Lombard General Insurance, India’s leading private sector non-life insurance company, is entering into a Bancassurance tie-up with AU Small Finance Bank....

Finance21 hours ago

Crypto’s tipping point

Chris George, Senior VP of Product at Somo argues that Crypto needs to improve its scalability to be taken seriously Cryptocurrencies are...

Business4 days ago

Why Procurement is key in delivering your ESG strategy

By Edward Cox, Principal at Efficio Consulting   Environmental, social, and governance (ESG) has shifted from a niche to a...

Finance4 days ago

Skedadle to change the game for advertising with Currencycloud partnership

Currencycloud, the experts simplifying business in a multi-currency world, has partnered with Scottish start-up app Skedadle to provide its users...

Finance4 days ago

How financial services organisations can harness the power of low-code/no-code

By Joman Kwong, Strategic Solutions Manager, Financial, at Laserfiche   The UK’s erratic economy, and its spiralling cost-of-living crisis, have...

Finance4 days ago

SaaScada Top Five Predictions for 2023

From BNPL for business, to sustainability and financial inclusion, 2023 is going to be a year of change as the...

Business6 days ago

Hidden channel costs: how to find and tackle them

By Mark Wass, Strategic Sales Director, UK and North EMEA at CloudBlue     Growth for businesses will always be a...

Finance6 days ago

Is your business ready for finance automation?

Mari-Frances Bentvelzen, Business Head and General Manager of Global SMB at SAP Concur   As managers continue to drive their...

Top 106 days ago

The power of a proactive customer service

By Delia Pedersoli, COO, MultiPay   2023 is shaping up to be another challenging period for B2C businesses. While the...

Business6 days ago

Automation nation: Liberating workers from desks, data entry and the doldrums

Gert-Jan Wijman, VP of EMEA at Celigo.   Just when businesses thought the tough times were over, even more challenges...

News6 days ago

Protean and Fino Payments Bank tie-up to expand PAN card issuance services in India

Fino Payments Bank has tied up with Protean eGov Technologies (formerly NSDL e-Governance Infrastructure Limited), a market leader in universal,...

Business6 days ago

What is the True Cost of SMS Phishing?

Gemma Staite, Threat Analytics Lead   Cybercriminals will recycle attack strategies for as long as they are effective. In Fraud...

Technology7 days ago

Digital Asset Management (DAM) To Transform Enterprise Brand Management

Alexander Rich, Co-founder and CEO – Desygner    Rapid digital transformation fuelled by the pandemic has undoubtedly proven beneficial to...

Finance7 days ago

Cost of living: How to identify vulnerable customers

Ellie Engley is account director at REaD Group   In the current climate, the cost of living crisis is a...

Banking7 days ago

Is traditional business banking the best option for SME finance squeezes?

Airto Vienola, CEO, AREX Markets  The pressures facing business and personal finances alike have been well documented. Stories are now starting...

Business7 days ago

Breaking down communications silos to streamline the customer experience

Dave Tidwell, Head of Technical Pre-sales, DigitalWell   The pandemic has, without doubt, moved the goalposts when it comes to...

Business7 days ago

How growth can be a big challenge when a business becomes multiple entities

By Paul Sparkes, Commercial Director of award-winning accounting software developer, iplicit. Organisations don’t just grow in size – they also...

Trending