• Lumiera
  • Posts
  • 🔆 Digital Diplomats & Biotech AI: Reflections on AI Use Cases

🔆 Digital Diplomats & Biotech AI: Reflections on AI Use Cases

Team Lumiera comments on Moderna partnership with OpenAI, the AI-generated foreign ministry spokesperson from Ukraine, and more

🗞️ Issue 17 // ⏱️ Read Time: 7 min

Hello 👋

In this week's newsletter

A deep dive into two relevant, and very different, AI use cases. Complete with comments from the Lumiera team, covering important aspects for leaders to look out for. Perfect conversation starters for your next team meeting and interesting case studies for the use of AI:

Lumiera Perspectives on Big Tech News

This week, we have chosen to bring your attention to two events in the AI ecosystem and share our team’s different perspectives.

Moderna’s (MRNA) objective was to achieve 100% adoption and proficiency of generative AI by all its employees with access to digital solutions in six months.

Emma’s thoughts on the Moderna x OpenAI partnership:

It’s not surprising that this type of partnership is announced by a company like Moderna. Innovation is part of the DNA of the pharmaceutical and biotechnology industry, and the level of competition is high, meaning that there is a need to move fast if you want to stay relevant. The company has a track record of rapid adoption in periods of change: In the pandemic, it was one of the first to release a COVID-19 vaccine approved by US authorities for emergency use.

AI solutions almost never create added value on their own, and Moderna’s leaders are mature enough to look beyond the hype of AI. Instead, a core product or a service that can be enhanced with AI is the way to go. To build substantial value with AI and drive real change, a significant mindset shift is required: AI must be integrated into the core strategy of the company, covering the goals permeating its core operations, services, and products.

Moderna announced that OpenAI’s product would be deployed across the organisation, “from legal to research to manufacturing”. Once again, this is an example of cross-disciplinary integration. AI trickles through the whole company, recognising the importance of the full workforce and how it augments their ways of working to deliver high-quality results. Simply put, Moderna has put into practice what is mentioned in this Forbes article: “leading a team or organisation in the era of artificial intelligence doesn’t just call for technical chops — what is necessary is business sense, empathy, foresight, and ethics.”

In my opinion, this is how real innovation is created: Giving people across the company, who know their work well, the opportunity and autonomy to find new solutions to old problems.

Allegra’s thoughts on the Moderna x OpenAI partnership:

The adoption of AI across Moderna is impressive, especially given the amount of employees (approximately 5600 people) and the complex nature of the data they work with. These variables equate to many more considerations when promoting tool usage, including usage guardrails, data security, and integrations into existing workflows, to name a few.

The success of their initiative stems from two major factors:

  • First is the custom transformation program they put into action, combining individual, collective, and structural change management initiatives. Moderna was truly dedicated to fostering a culture of learning at all levels of the company and ensured every single employee was incentivised to engage in the journey. Technology is only part of the story, the importance of a people and process-focused approach cannot be understated.

  • The second critical factor to Moderna’s success with AI was their investment in their tech stack and data platform over the last decade.

When implementing an enterprise generative AI solution, it’s expected to spend some time fine-tuning the base model on the organisation's data to ensure that output is more domain-specific. At Moderna, they’d need to train ChatGPT on the terminology specific to mRNA medicines. This process often involves collecting training data to ensure good performance. Additionally, they would have had to determine how ChatGPT Enterprise would communicate with their other applications and back-end systems. This means defining the data ingestion process enabling ChatGPT Enterprise to securely consume and update necessary data while adhering to their unique standards within a highly-regulated industry.

This is quite the challenge when you account for the sheer volume and variety of data Moderna works with as they build the industry's leading mRNA technology. Their long-term commitment to foundational data platform work has clearly paid off and is something other companies looking to adopt AI at scale should pay close attention to.

🎙️ Ukraine unveils AI-generated foreign ministry spokesperson to read its human-written statements.

Dmytro Kuleba, the Ukrainian foreign minister, said Victoria was created to save diplomats’ time. To avoid fakes, the AI read statements will be accompanied by a QR code linking them to text versions on the ministry's website.

Sarah’s thoughts on Ukraine’s AI spokesperson:

The first thing I thought of was how representation of Afro-Europeans in various ministries of foreign affairs around Europe is astoundingly low, and so seeing this digital AI diplomat left me with mixed feelings.

On one hand, we know that representation matters. So a young black Ukrainian may see the digital AI diplomat and think: “Hey, I want to do that! I can do that.” On the other hand, representation is not enough. If there are no serious policies put in place to increase diversity (both regarding the recruitment and retention of personnel) in the diplomatic service - what is actually happening here?

I have questions, the main one right now being: to which extent does using AI allow government agencies or organisations to rebrand themselves in ways that do not reflect reality?

It is imperative that government agencies and ministries stay relevant and up to date with technological development. But this cannot stop at small steps towards efficiency - they must be accompanied with relevant and robust policies to make sure that use is responsible and transparent.

Beyond this, the public sector needs to show serious consideration for AI’s role in a broader context with many factors at play - geopolitics, conflict, climate change… not a small feat, and I look forward to seeing leaders in the public sector rise to the occasion.

Allegra’s thoughts on Ukraine’s AI spokesperson:

The details behind the technology powering Ukraine's AI spokesperson, Victoria Shi, aren't entirely public. However, based on news reports, we’re likely seeing a combination of existing technologies for computer graphics, text-to-speech, and machine learning.

  • Voice & Visual Data: Shi's appearance and voice are modeled on a real person: Rosalie Nombre, a singer and former contestant on Ukraine's version of The Bachelor reality show, who agreed to the use of her voice and likeness "pro bono".

    • High-quality voice recordings of Nombre would have been used to train the AI model to generate natural-sounding speech. These voice recordings capture the pronunciation, intonation, and other vocal characteristics that the AI spokesperson will emulate.

    • Next, visual data is needed. This could include images and videos of Nombre, serving as the basis for the AI-generated spokesperson's appearance. The visual data helps in modeling facial expressions, gestures, and overall appearance.

  • Text Data: To generate spoken content, you need text data. This text can include scripted statements, speeches, announcements, or responses to specific questions. The text serves as the input for the AI system to generate the spokesperson's voiceover. As stated previously, The Ministry of Foreign Affairs clarifies that while Shi delivers the messages, the content itself is written by humans.

  • Text-to-speech conversion: Shi's voice is likely generated using text-to-speech software, which can create realistic speech from written text. There are many options for this on the market, so the requirements for Ukraine’s particular use case would have been closely considered including factors like voice quality, multi-lingual support, and integration availability.

What we are excited about:

Lumiera’s Q2’24 AI Must Knows:

A thoughtfully curated, comprehensive, FREE set of resources with everything you need to get caught up on the state of AI at this moment:

✅ The best free or low-cost online AI courses

 Recent industry reports

  Impactful research papers

 AI jargon dictionary

The Luganda Neural Text-to-Speech (LNTS) system is a testament to the possibilities of leveraging technology to break barriers and promote inclusivity.

Developed by researcher Ronald Kizito, the LNTS system was designed to cater to Luganda speakers, particularly those facing visual challenges. It represents a significant step toward promoting health and accessibility within Luganda-speaking communities.

Until next time.

Emma - Business Strategist
Sarah - Policy Specialist
Allegra - Data Specialist

Lumiera has gathered the brightest people from the technology and policy sectors to give you top-quality advice so you can navigate the new AI Era.

Follow the carefully curated Lumiera podcast playlist to stay informed and challenged on all things AI.

What did you think of today's newsletter?

Login or Subscribe to participate in polls.