[Exclusive Interview] This Software Development Company Is Determined To Eradicate Data Bias For Optimizing AI Usage

Recently, we interacted with Mr. Philip Miller, Customer Success Manager, Progress – a software innovation company that enables customers to launch and deploy powerful products for scaling their operations and services.

Here are the interview highlights:

1. What is data bias and how does it affect the accuracy and fairness of data-driven decision making?

Data bias in artificial intelligence (AI) is a critical issue that warrants our attention. It refers to the presence of errors or distortions in data that can lead to unfair decisions that perpetuate discrimination and inequity. It can also lead to security and governance risks, reputational damage, lost customer trust and ethical exposure. Organisations must proactively ensure diverse and representative datasets, rigorous evaluation methods and ongoing monitoring to mitigate data bias. By doing so, they can harness the power of data-driven technologies while promoting accuracy, fairness and ethical practices in decision making. Also, addressing data bias has a broader societal impact, contributing to a fair and inclusive society.

2. Can you describe the methodology and objectives of your recent data bias survey and why the company conducted this research?

Progress partnered with the independent UK-based research firm, Insight Avenue, to conduct an extensive study on AI data bias to understand the magnitude of the problem and see what businesses are doing to address it. The study aimed to provide insights that would help companies improve their digital ethics strategy and proactively tackle data bias issues.

Mr. Philip Miller, Customer Success Manager, Progress

The research team conducted 640 interviews with business and IT professionals at the director level and above, specifically targeting individuals who use data for decision-making and are either currently using or planning to implement AI and ML solutions. The survey spanned 12 regions, including Europe, Asia, Latin America and the United States, ensuring a global perspective on data bias awareness and mitigation efforts.

According to the results, an overwhelming majority of respondents, 78 per cent, expressed concerns about the growing impact of data bias as the use of AI and ML increases. Moreover, 65 per cent admitted the existence of data bias within their own organisations.  However, the study also revealed that many companies need to do more to address data bias, with 77 per cent acknowledging the need for further action.

Interestingly, the study found that only 13 per cent of organisations are currently actively addressing data bias and have established an ongoing evaluation process to identify and mitigate biases. It suggests that there is still a significant gap between awareness and effective action. Lack of awareness and understanding of biases was cited by 51 per cent of respondents as a barrier to addressing data bias, indicating the need for education and increased knowledge in this area.

3. What is the situation in India of AI and Machine Learning systems?

The recent survey conducted by Progress in India throws light on the prevalence of AI and machine learning systems. The findings indicated that 66 per cent of the respondents anticipate an increased reliance on AI in the next five years. However, a higher percentage, 78 per cent express concerns about the growing problem of data bias as their dependence on AI/ML systems grows. These concerns are consistent across different regions in India. Within organisations, 57 per cent of business decision makers express concerns about data bias, and 65 per cent believe that data bias already exists within their organisations. The impact of data bias is evident in various decision-making areas, including finance, IT/digital, operations and customer acquisition. This is important, because biased decisions can lead to negative consequences for customer experience, cause potential legal issues and hinder inclusion and diversity efforts.

The survey also showed that most businesses are still in the early stages of understanding and tackling this issue. Effective measures to combat data bias include supporting education and training, ensuring transparency and traceability of algorithms and data, allocating more time for model training, and using tools to identify bias within data sets. However, 77 per cent of respondents believe that their organisations need to do more to comprehend data bias. Technology and tools are deemed the most urgent to combat data bias (65 per cent), followed by additional training (59 per cent) and strategic adjustments (49 per cent).

4. Can you share examples of the negative impact of AI data bias on people?

Biases can arise in various areas beyond traditional data sets and addressing them requires a holistic approach that combines technical considerations with diverse perspectives and inclusive practices. The consequences of biased AI data can be severe in different sectors. For example, in the retail industry, flawed hiring algorithms that favour male candidates for technology roles perpetuate gender disparities. In finance, flawed AI tools may discriminate against certain zip codes, resulting in qualified loan candidates being incorrectly rejected, thereby exacerbating socioeconomic inequalities. In healthcare, using AI to determine healthcare eligibility can deny individuals the proper care they deserve, leading to adverse medical outcomes and reinforcing healthcare disparities.

It’s worth noting that the development of AI systems is predominantly driven by a small pool of technical experts, often with limited diversity in their perspectives. This homogeneity can introduce inaccuracies and biases into the data. By recognising these challenges and working towards diverse and inclusive AI development practices, we can mitigate data bias and strive for fairer and more ethical AI systems that benefit everyone.

5. How is Progress using Generative AI to transform customer service and experience?

Progress leverages generative AI within its DX technology to revolutionise customer service and elevate the overall customer experience. With generative AI, Progress customers can get a deeper understanding of their customers, which allows them to deliver highly personalized experiences, at scale, that are more engaging and relevant. 

Also, generative AI enhances content generation and delivery through tools like Sitefinity CMS. By analysing customer interactions across channels, generative AI assists in creating highly personalised content that resonates with the target audience. This automation improves efficiency and ensures timely and relevant content delivery across various devices, enhancing customer engagement and satisfaction.

Furthermore, generative AI powers conversational interfaces for enhanced customer service. Through Native Chat businesses can leverage generative AI to build chatbots and streamline support workflows. Automation flows driven by generative AI enable the resolution of issues either proactively or reactively, resulting in faster response times and improved customer satisfaction. By providing adaptive technical support and personalised recommendations, generative AI enhances the overall customer service experience.

6. What are some of the challenges organisations and professionals using Generative AI face? 

One of the important challenges is the potential bias in generated content. Generative AI models learn from existing data, which can introduce biases present in the training data. It is important for organisations to carefully arrange and review the training data to ensure that the generated content does not construct any biases. By actively monitoring and refining the training data, businesses can mitigate the risk of biased content generation.

Another challenge is the need for human oversight and control. While Generative AI can automate and streamline content generation, human oversight is still necessary to ensure quality and brand consistency. Professionals should actively participate in the training process, monitoring and fine-tuning the models to align with their specific goals and requirements. This collaborative approach allows organisations to strike a balance between automation and human judgement, resulting in high-quality generated content.

Data privacy and security are other important considerations. Generative AI models require substantial amounts of data for training, and organisations must handle this data responsibly. Compliance with data protection regulations and implementing robust security measures are crucial to safeguarding sensitive customer information. By prioritising data privacy and security, businesses can build trust with their customers and ensure the ethical use of Generative AI technology.

Finally, organisations should be mindful of the ethical implications of using Generative AI. The potential for deepfakes or misleading content poses ethical concerns, and businesses must employ Generative AI responsibly and transparently. Clearly communicating the use of Generative AI-generated content to customers and stakeholders helps build trust and ensures ethical marketing practices.

By acknowledging and addressing these challenges, organisations and professionals can harness the full potential of Generative AI in their marketing efforts. Through a careful curation of training data, human oversight, data privacy measures and ethical considerations, businesses can leverage Generative AI to create impactful and engaging content while upholding their values and maintaining customer trust.

7. What is the future of AI?  Do you think organisations will trust it more or issues like data bias will restrict its full use?

AI has a wide range of uses in businesses, such as AI in education, healthcare industries, transportation, finance, human resources and many more. As technology advances, the world will see more start-ups, different business applications and consumer uses. AI will dramatically remake the entire ecosystem and play a major role in transforming various industries. AI is developing at a rapid pace. However, one must address the concerns surrounding data bias to ensure the full potential of AI. .  Today’s organisations are increasingly becoming aware of unbiased and ethical AI systems. By executing robust data governance practices, ongoing monitoring and transparent algorithms, organisations can mitigate biases and build trust in AI. Organisations will gradually embrace AI with more confidence and leverage its capabilities to drive impactful outcomes.

8. What ethical considerations should data scientists and businesses keep in mind when working with biased data?

Big data is growing every day and different sectors and government organisations are using it on a large scale. The amount of data generated by businesses in every sector is unprecedented. Therefore, when it comes to biased data, data scientists must prioritise a few ethical considerations. First, it is important to acknowledge and identify biases in the data, ensuring there should be proper transparency throughout the process. Data scientists should strive to diversify the dataset, actively seeking out representative samples to mitigate bias. Also, every organisation should establish a strict policy for handling biased data, promoting fairness and accountability. Regular audits and monitoring can detect and rectify biases that may arise during data analysis. Finally, collaboration with the diverse teams and merging different perspectives can aid in addressing biases effectively. By adopting these measures, data scientists and organisations can work towards minimising the impact of biases and fostering a more ethical approach to data bias.

9. Do you have any career advice for those who are planning their career in data science?

I have a few key pieces of career advice for those planning their career in data science. First and foremost, focus on building a strong foundation in mathematics, statistics and programming. These skills are fundamental to excel in the field of data science. Also, develop a curiosity-driven mindset and continuously expand your knowledge by staying updated with the latest industry trends and advancements. Practical experience is invaluable, so seek out opportunities to work on real-world data projects or participate in internships to gain hands-on experience. Networking and collaboration are also crucial—engage with data science communities, attend industry events and connect with professionals to build a strong network. Also, never stop learning and be open to continuous learning and upskilling as the field of data science evolves rapidly. By following these career advice tips, one can pave a successful path in the dynamic and exciting field of data science.

I’d add that as a good AI developer, engineer, or architect, you should be looking at other disciplines to inform and improve the outcomes of your AI.  Look towards psychology, philosophy, history, biology, and anthropology to name a few; to help answer some of the fundamental questions you might have about your AI and help to align it towards goals that align with ours.  This will go a long way in making sure that some of the fears around the technology are mitigated, as well as helping to address future regulatory pressures.

Image Source