Exploring the future of AI with Phelim Bradley, Chief Executive Officer, Prolific

Ahead of World Summit AI (11th-12th October 2023, Taets Art & Event Park, Amsterdam), we asked Phelim Bradley, Chief Executive Officer, Prolific, his thoughts on the future of AI.

As an expert in the field, what critical challenges do you believe the AI community needs to address to ensure responsible & and ethical AI deployment?

The people behind your data matter. We need to make sure that we have a diverse workforce providing the human intelligence that makes artificial intelligence better. It’s critical we make sure that people are representative of the population if we are to have accurate and unbiased AI. And also that we are protecting human dignity as we develop these tools. It’s important that we develop AI systems that augment people and don't replace them. The goal should be making humans more efficient or more productive. That’s key for developing a prosperous future.

How has AI impacted your specific field of expertise, and what transformative changes do you foresee in the near future?
  

There’s been positive and more challenging changes, and I’ll start with the more challenging. We’ve seen an increase in the use of AI by participants in online research, which is reducing data quality. Essentially, people are using AI to augment their own responses in a way that was not intended by the researcher. On the good side, there's massive value in research summarisation and understanding what's been done in the field already. Being able to summarise and explain research means you no longer have to be an expert to understand academic papers - you can get AI to explain it to you.

 

How do you envision AI shaping various industries, and what advice would you give to businesses seeking to integrate AI into their operations?

  

The most diverse use case of AI is going to be augmenting people's day-to-day use - primarily as a writing and communication tool. In terms of advice, look first to augment people's productivity rather than thinking of AI as a replacement for human intelligence. Understand the source of the data used to train these tools, making sure it is sufficient for your use case. And give clear guardrails for how you want people to use AI - how much private data can be uploaded? What training clarity is available on how best to use these tools?

 

In your opinion, what opportunities and challenges does AI present for job markets and workforce development worldwide?

Roles that leverage AI and human collaboration will rise. The earliest form of that would be Prompt Engineers, or people who are experts at maximising the outputs of AI tools. Again, it’s important that it is a productivity enhancer, not a human replacement. Related to that, helping people do their job effectively alongside AI is crucial. In certain industries - and knowledge workers are particularly affected by this - there’s a need to re-skill people, training them in how to utilise AI to do their jobs more effectively.

 

Can you share an example of an AI application or project that has personally impressed you, and explain why it stands out?

  

Hume AI has worked on gathering speech and emotion data from people, designing a toolkit that understands human emotional expression. They did a very cool demo recently looking at the song “Used To Be Young” by Miley Cyrus, mapping her emotional expressions throughout the video to predict her pain, joy, fear, relief and so on. That’s a powerful example because it uses AI to understand human behaviour beyond just Natural Language Processing.

 

 

 

What measures do you believe should be taken to bridge the AI research gap between developed and developing nations to ensure equitable technological progress?

  

To address the AI research gap between developed and developing nations, two key strategies are pivotal: Access to Quality Data and Inclusive Research. First, equalising access to datasets is essential - open science brings more people into the community. Open-source initiatives can democratise this access, empowering researchers globally. Second, it's crucial to encourage an inclusive research ecosystem by involving participants and researchers from diverse and often marginalised backgrounds, including those in developing countries. These two facets combined can democratise AI knowledge and technology, fostering a more universally ethical and beneficial progress in AI.

 

What 2 people do you admire most in the world of AI in terms of their work?
  

I find Clément Delangue and the Huggingface founders admirable - notably their commitment to developing AI in an open and community-driven manner, which I believe will also support the points made in the previous question. And I find Mustafa Suleyman from Inflection impressive too for his commitment to safe AI development.

 

What advice would you give to aspiring AI researchers and enthusiasts who want to make a positive impact in the field?
  

There’s three key points. Strive for interdisciplinary, collaborative and community-driven projects - ones that prioritise the human being at the centre of both AI development and AI use. Consider ethics as the first priority. And build things that make humans better - things that are not intended to replace people.

 

If you could solve any global problem in the world with AI, what would it be and why?

 

I did my PhD in the genetics of bacteria, and developing tools to analyse the DNA of bacteria to detect drug resistance and solve the problem of anti-microbial resistance. I am going to be watching AI's application in biology and genetics with interest. There’s some pretty meaty problems there that could be solved with AI use. Although Large Language Models are focused on human language, the DNA behind genetics is a language too.

 

What inspired you to participate in this AI summit as a speaker, and what message do you hope to convey to the audience?
  

The very heart of my talk, and why we decided to come to World Summit AI, is that everyone is aware of the fact that there is a lot of data powering the development of AI tools. Therefore increasingly, there is the need for scientifically valid data sets. Human intelligence is at the heart of developing Artificial Intelligence, so my message to the summit is that the people behind the data matter. Let’s build AI together - in an inclusive, collaborative and open manner.