Shutterstock’s Senior Director of Artificial Intelligence and Data Science explains how she’s building a team to develop ethical AI systems.
Artificial intelligence is a relatively young technology. Still, it’s evolving quickly.
You may be familiar with AI from the buzz in the marketing world. After all, many strategists and CMOs predict it to be the next big technological solution for the industry. Yet, we wonder about the advantages AI brings beyond the scientific world.
Can AI make us better creative thinkers and problem solvers? If so, what ethical parameters do algorithms need to follow, in order to be fully representative of the human experience? Why are these ethics so critical to the development of AI in the first place?
Shutterstock’s team of data scientists and engineers are working toward developing AI solutions that inspire creativity.
Alessandra Sala, Shutterstock’s Senior Director of Artificial Intelligence and Data Science, is an innovator who’s leading a diverse team of international scientists to build systems that inspire and nurture creative thinking.
Sala says that the blend of both human ingenuity and data-driven exploration will take creativity to new heights.
Here, Sala talks about her experiences and priorities in creating ethical AI . . . and how her team is best equipped to bring the most well-rounded creative solutions to the global marketplace.
Shutterstock: Can artificial intelligence make us better creative thinkers?
Alessandra Sala: There are several aspects we ought to consider when we discuss creativity in the context of AI. Let’s say we merely focus on the question, “Can AI generate an image?” The answer is yes. Today’s generative AI has proven its ability to generate diverse and realistic concepts, along with fantasy.
Specifically, we have seen several different generative AI systems that create imagery that’s based off of text prompts. However, are those generations equal or better than human creativity?
It’s important to understand that these types of systems are replicating what they’ve seen during their development. They are technically learning from a massive data set of many examples. They match text to images and vice-versa.
So, they recombine their learned concepts, following the instructions of a human-generated text prompt. Because of that, I would argue that creativity resides in the human who is able to conceive a creative text prompt to feed the AI system.
On the other hand, I’m mostly interested in a slightly different take on this discussion. Can artificial intelligence inspire human creativity?
AI is a brilliant tool to help people trigger their own creative thinking processes. The scientific community has conducted several experiments between artists and AI systems. When an artist leverages generative AI technology, innovative outcomes can reach new heights.
I strongly believe AI can amplify human creativity in new ways. AI’s power is the ability to quickly explore the “infinite” space of possibilities (which humans can’t) and synthesize a smaller set that can be curated and contextualized with human common sense.
SSTK: How will AI be useful to creative thinkers in the future?
AS: We are at the dawn of a new creativity era. AI will play an important role in this journey. Currently, there are few approaches to connect these technologies into our workflows, but we are early in the experimentation phase. The more we integrate these new capabilities, the better we understand the new opportunities we’ll be able to develop in years to come.
As we build this future, our attention should be laser-focused on developing technology which will not create unintended consequences, such as infringing human rights or fundamental freedom.
Therefore, we need to constantly ask ourselves:
- How can we operate in a world that we don’t fully understand yet?
- How can we build systems that respect human rights and principles?
- How can we equitably enable humans and machines to collaborate in new ways?
These questions do not have predefined or standard answers, but they should be considered in each research and innovation development to guide our thinking. Those questions should be leveraged as guardrails to audit new ideas. This will help us to develop equitable technology.
I care about this so much, and I am also actively supporting the UNESCO agenda around Ethical AI. In November 2021, I was the only industry representative at UNESCO. At that time, I presided over the press conference for the launch of the Recommendation of Ethics in AI.
SSTK: Can you detail what Shutterstock is doing to advance creative AI technology?
AS: We fundamentally believe that superior creativity comes from inclusively connecting diverse perspectives, creative thinking, and novel technology. Shutterstock is working on cutting-edge AI technology in the creative space.
Shutterstock’s AI team is also built with diversity as a core principle. Our AI scientists are not only technically brilliant, but they bring diversity in terms of cultural, educational, and gender representation.
One of our data science teams has already reached a record 40/60 gender balance, which is largely ahead of industry standards. We have representation from so many countries, across Europe, Africa, Asia, and North America. Our team has a rich educational diversity too, with experts in computer science, engineering, astronomy, and neuroscience.
As our team is made up of people from various backgrounds with various skill sets, we constantly challenge the status quo and strive for inclusive design. In order to create AI that is ethical, it needs to understand data through the eyes of diverse perspectives.
This diversity is critical. It ensures our systems consider equitable outcomes, as we want to appeal to Shutterstock’s global contributors and user base.
All people deserve to see themselves represented in the solutions we’re building for them.
SSTK: Alessandra, you have spoken about the importance of diversity in AI. As Global President of Women in AI, which are the biggest barriers which prevent the growth of more women in AI and STEAM?
AS: The lack of women in technology is an issue that, I believe, has been approached from the wrong perspective. For years, we have kept hoping to hire more women. This has been done without putting any effort in growing women’s interest in tech, nor supporting smarter education programs. The right educational programs could help more women to enroll in STEAM (science, technology, engineering, arts, and math) studies.
On the other hand, we have focused on gender index. This is critically important, but it only measures the symptoms, instead of the causes.
Women in AI is a nonprofit do-tank working toward inclusive AI that benefits global society. We are a community-driven initiative bringing empowerment, knowledge, and a plethora of educational programs. These efforts are meant to equip more women to enter the AI field as technical experts, while supporting their career growth.
Understanding cultural biases, as well as the role that women can play in AI, is a shared leadership responsibility. It will help our industry to build AI-powered products that demand fair and equitable outcomes for everyone.
Ready to unleash your creativity with the power of AI? Just type a description of something to get started.
AI Image Generator: Create stunning visuals instantly
License this cover image via Jorm S.