Gaurav Pathak
Vice President of Product Management at Informatica
For businesses to successfully implement AI, high-quality data is essential. With the success of generative AI (Gen AI) being built on data, organisations implementing Gen AI initiatives need to ensure that it is unbiased and secure. To succeed in this, businesses can benefit from investing in data management solutions to improve the quality of their data.
This month, we speak with Gaurav Pathak, Vice President of Product Management at Informatica, about how Informatica is moving towards ethical AI and what businesses can do to achieve fast AI in a responsible way.
Informatica’s recent CDO survey found that the quality of data is the top data-related obstacle to implementing AI. Why do you think this is the case?
There’s an old IT proverb – garbage in, garbage out – or simply, you are what you eat. Just as high-performance athletes can’t expect podium-worthy results on a diet of chips and soda, the output of LLMs and AI solutions are directly influenced by the quality of data that nourishes them.
The reality is – at the enterprise level – everyone claims to be ready for AI, but is their data? Hybrid, multi-cloud, multi-vendor data estates remain largely fragmented and incredibly complex, disparate, burgeoning ecosystems of structured, semi-structured and unstructured data that, according to our recent survey results, are only expected to worsen in the months ahead due to demands from Gen AI projects.
This is death by data: a corporate paralysis that comes from organisations’ inability to properly manage their most important asset. This lack of trust and confidence is justified, given that organisations have not focused on data quality and governance as much as they should have.
And now Gen AI is bringing new but similar requirements like identifying and correcting/removing inaccurate, incomplete, or irrelevant sections from unstructured data, removing confidential and sensitive information about customers and employees, excluding content based on freshness/data ranges and more in addition to ensuring structured data is in good shape. Realising tangible ROI from Gen AI requires high-quality, secure, and well-governed data and the time to act is now.
How can businesses best accelerate their AI use?
Get their data houses in order. There was hardly an enterprise that wasn’t eager to board the Gen AI hype train last year, but without tracks built on sound data, business leaders flirt with a catastrophic – and costly – derailment in 2024 as they move from AI exploration to actual implementation.
We know data fragmentation and complexity continue to be a real struggle for many organisations, regardless of geography or vertical. Our recent survey found that four in 10 data leaders struggle to balance 1,000+ data sources, and eight in 10 anticipate this number will only increase this year.
For organisations to bring calm to this data chaos and accelerate their AI adoption, they need to put the cart back behind the horse and invest in the integrated data management solutions that can deliver trusted, secure data, reduce costs and drive measurable growth. As our CEO Amit Walia relayed to Jon Fortt of CNBC during recent earnings conversations, “Gen AI has made data management a business process; without a good, clean data layer, there isn’t Gen AI.”
Why is ethical AI so important and how can businesses ensure responsible AI operations?
Accuracy, bias and hallucinations are all valid, well-documented concerns when it comes to AI, particularly around LLMs and the Gen AI solutions they power. Without implicit trust in the output, these innovations are little more than advanced party tricks. Unsurprisingly, nearly 40% of data leaders we surveyed cite AI ethics as a key hurdle to AI adoption.
AI and Gen AI must be never considered a ‘set it and forget it’ tool – human oversight to continuously refine and fine tune models must be built into the lifecycle. Regular and ongoing feedback loops are paramount to limit the impacts of this unwelcomed trinity.
From a data perspective, responsible AI operations should be anchored around two key ingredients – privacy and governance. Our survey confirmed data privacy and protection are top challenges to AI and Gen AI adoption, and they’re also the data management capabilities most in demand by leaders to support their data strategies.
What are some AI innovations that Informatica is working toward at the moment?
Our journey with AI predates the current hype era by three or four years. In 2018, after we completely revamped our business trajectory and product suite to be cloud only, we unveiled CLAIRE – the industry’s first metadata-driven, AI-powered engine for data management.
CLAIRE, which automates and scales thousands of intensive tasks across the enterprise, currently processes 86+ trillion monthly cloud transactions, a volume that increases at a rate of about 60% year-over-year. It’s the backbone to the Intelligent Data Management Cloud, our holistic data management platform, and each of our bespoke products and solutions.
Last year, we announced a suite of new AI copilot capabilities and CLAIRE GPT, the next generation, Gen AI-powered version that leverages the simplicity of a natural language-based interface, which will dramatically accelerate how enterprises ingest, process, manage, explore and store data. With CLAIRE GPT, which will be GA in the first half of the year, customers will enjoy dramatically improved productivity, reduced costs and increased self-service for users of all technical abilities.
Additionally, we invested significant resources to advance many of our strategic partnerships with other industry leaders throughout 2023, such as AWS, Microsoft, Snowflake and MongoDB, with an eye to expand this ecosystem further in 2024. These included deeper integrations that will empower joint customers to gain a better grasp of their data landscapes in preparation for AI and Gen AI workloads.
Moving forward, how will Informatica ensure that it will work toward ethical AI?
Many Informatica products and services feature technology that enables our users to process information with an increasing degree of autonomy. At Informatica, we understand the profound impact of AI that makes this automation possible, and we guide our AI development with an ethical, responsible, and comprehensive set of principles.
These principles are designed to ensure that the AI technologies we create and deploy are developed and used in a way that respects human rights, contributes to societal benefits, upholds privacy and security, prioritises transparency and explainability and strives for inclusivity and diversity.
We aim to democratise AI, providing tools that are accessible to all users regardless of technical expertise. Our commitment also extends to not designing AI for deployment in ways that can potentially cause harm or undermine the values that we stand for.
We understand that the field of AI is rapidly evolving, and thus, we will reassess and update these principles to keep pace with technological advancements and emerging ethical considerations. We firmly believe that by adhering to these principles, we can drive progress while ensuring the responsible use of AI.
What does the future look like when it comes to ethical AI?
Often, we think of external forces as the main obstacle to growth and transformation, but our survey findings reiterated an uncomfortable truth – internal resistance due to bad quality and ungoverned data affects AI and Gen AI priorities the most. Nearly all data leaders we canvassed (98%) admitted organisational malaise holds back their data strategies, with the primary culprits being lack of leadership support, inability to justify ROI and lack of alignment between teams.
But our findings also reinforced what we already knew – there is an indisputable, intrinsic connection between secure, sound data management and successful AI, and an investment in these advanced technologies continues to fuel mutual investment in data management capabilities like privacy, quality and integration. It’s little surprise that nearly eight in 10 data chiefs predict an increase in their data investments this year while all plan to specifically invest in data management solutions.
We believe for today’s enterprises to be tomorrow’s market leaders, they must sit at the intersection of cloud, data and AI – which is exactly where you’ll find Informatica.
************
Make sure you check out the latest edition of AI Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
************
AI Magazine is a BizClik brand
