(EMEA) VIDEO | Clean Data Is a Must to Limit Hallucinations — Ipsen CDO

Ipsen CDO Andrea Pezzetta speaks about organizational strategy around buy versus build, responsible adoption of generative AI, the need for transparent partnership, and future opportunities in the pharma sector with AI.

Andrea Pezzetta, Chief Data Officer at Ipsen, speaks with Huma.Ai CEO and Co-founder Lana Feng, in a video interview about organizational strategy around buy versus build, leveraging generative AI to derive quick value with use cases, responsible adoption of generative AI, the need for transparent partnership, and future opportunities in the pharma sector with AI.

Ipsen is a French biopharmaceutical company headquartered in Paris.

At the onset, Pezzetta opines that every company is different in terms of strategy. She believes that the size of the organization impacts what can or cannot be done. Commenting on the build versus buy aspect, Pezzetta prefers a mixed model to determine where to build and buy to bring the right drugs to the market for patients quickly.

Typically, she says, smaller organizations tend to buy versus build because it is simpler to put things into effect.

Also Read
(EMEA) VIDEO | We Need People Who Really Understand Data —  Ipsen CDO
(EMEA) VIDEO | Clean Data Is a Must to Limit Hallucinations — Ipsen CDO

When asked about the CDO’s perspective on forming startups, Pezzetta maintains that in the health industry patients must come first and privacy is a major concern. She states that startups must ensure that improper data is not exposed and there is no black box.

As a CDO, Pezzetta is particular about putting integrity first when it comes to looking at the buy option. She is concerned about the data going in and ensuring that data is protected.

Taking patient data for example, Pezzetta insists on doing the right thing for patients. Further, she reiterates that having a mixed model allows employees to build in-house, innovate, and grow skill sets.

Moving forward, Pezzetta discusses the need to focus on use cases that generate quick value with generative AI. She mentions that diversity and health equity are a challenge in the healthcare industry, and she would want generative AI to help with getting unbiased clinical trials.

Next, Pezzetta says that generative AI can be leveraged to speed up protocol design and accelerate drug discovery. Leveraging massive amounts of data with great tech goes hand-in-hand, she affirms.

Additionally, Pezzetta mentions the golden batch use case in the manufacturing sector. She states that it ensures attaining a standard level with processes and helping that along with Gen AI is critical for the industry.

It is imperative to understand what derives the best value for the organization and patients and use such use cases going forward, says Pezzetta. While she would love to employ generative AI across the board, the organization has yet to reach that point and has to test out different solutions to derive the best results.

Speaking on the responsible adoption of Gen AI, Pezzetta highlights the importance of having clean data. The biggest wish, she says, is to leverage AI to clean data and look for anomalies.

Privacy and security are potential challenges, adds Pezzetta. She asserts that the organization is working to put proper guardrails in place, and companies across the board must lay the framework regarding what good data should look like with Gen AI.

Pezzetta stresses understanding where to experiment, where to find value, and how to tie the value to deliver better products to patients.

Shedding light on hallucinations, she states that clean data is a must to limit hallucinations. Pezzetta recalls reading about “Letting AI rest,” and for this, the organization must ensure that the training data is clean, accurate, and unbiased.

Getting back to privacy and security, Pezzetta says that there have been numerous records of corporate data being exposed to the public because of people using public tools. She states that organizational CSOs, CEOs, and DPOs must come together and tackle this as a team by having the right rules in place.

Emphasizing the need for transparency, Pezzetta urges organizations to be fully aware of the partner companies and how the partnership can be beneficial.

Furthermore, she states that in the future, she would like to see the clinical trial space grow with AI, which would allow organizations to get closer to precision medicine. Further, Pezzetta mentions leveraging real-world data to the efficacy of drugs and genomic data to understand underrepresented diseases.

Concluding, she affirms that every pharma company would like to witness that progress and it is a huge opportunity to bring drugs to patients faster.

CDO Magazine appreciates Andrea Pezzetta for sharing her insights with our global community.

Executive Interviews

No stories found.
CDO Magazine