Sharktower’s Senior Product Manager says “Aye!” to Scotland’s AI Strategy
Resource Type: News | Published: 22nd Mar 2021 | By Team Sharktower
On Monday 22 March 2021, Scotland’s AI strategy launched, outlining a 3-track approach to support and develop AI-related initiatives that will contribute to Scotland’s economic, social and environmental future. The launch session included an introduction from the Chair of the Steering Committee, Kate Forbes MSP, along with panel discussions including spokespeople from various industries in Scotland as well as committees from Denmark, New York and New Zealand.
Scotland’s AI strategy focuses on ‘trustworthy, ethical and inclusive AI’ and has been created through a collaboration of the Scottish Government, The Data Lab: Scotland’s Innovation Centre for Data Science and AI. There has also been strong input from various Working Groups, each with representatives from Scottish-based businesses and organisations, including Sharktower’s Senior Product Manager, Mark Graham.
We asked Mark to tell us a bit more about the strategy and how he became involved.
Why do you think it’s so important that Scotland has an AI strategy?
“With more and more companies introducing AI to their products, it’s really important that Scotland has a strategy in place, with a vision and principles around AI to ensure it has a positive impact on society. There are risks that come with AI if not used in the right way and I think a key factor is trust.
There can be many negative connotations associated, so by having this strategy in place we can work together to mitigate these risks and build AI for good. I’m really excited by the concept of the Scottish AI Alliance which will bring experts from around Scotland together to build partnerships and work towards achieving the vision.”
How did you get involved with the AI strategy working groups?
It was actually through The Data Lab, Scotland’s Innovation Centre for Data Science and AI. They have supported the Scottish Government to develop the AI strategy and I applied to be involved with the working groups they were running. The group I was part of was called ‘Developing AI and AI-Enabled Products and Services’ and I was lucky enough to work alongside some amazing people from around the country.
How do you see the theme of ‘trustworthy, ethical and inclusive AI’ being applied in organisations such as Sharktower?
We’ve always aimed to build responsible AI, so we’re already doing most of it, but I think it’s great there’s now a national framework in place. For us at Sharktower, that means continuing to develop our 4 principles for building responsible AI, which are human-centered design, governance, monitoring performance and training data.
In terms of human-centered design, well, to add AI to software, we need to focus on user-centric design to ensure that the overall experience compliments the data we’re presenting. We regularly review our user personas to help understand user’s needs, experiences and goals along with outlining the types of behaviour we want to encourage. Requirements include privacy, transparency and security and we continually test and gather feedback so we can iterate the design and user experience.
For governance, we’ve adopted an Ethics Framework to support the development of our AI-enabled features and we also ensure the use of data and development of AI aligns with our core values that we live by in our working lives.
When it comes to monitoring performance, all AI needs close supervision using ongoing human monitoring and auditing of the performance of algorithms. Our data scientists have recently revisited and reviewed all of our models and we’re constantly looking to improve them.
Then finally there’s training data. As well as reviewing models and performance, the data sets that the models have been trained against need to be continually reviewed, as an example with Team Sentiment, we look at new project languages and definitions to ensure the model understands new project terminology.”
What role do you believe Sharktower is able to play in developing Scotland’s AI strategy going forward?
“Well I’d love for us to become part of Scotland’s AI Alliance, which will be created during Track 1 of the wider strategy. But in any case it’s important that Sharktower aligns with the strategy overall, and we’ll continue to give our support and expertise to the Scottish AI Playbook.”
Read Scotland’s AI strategy in full on the official website
Watch a demo of Sharktower’s Explainable AI features
In February 2021, Mark hosted a live walk-through introduction to explainable AI, and show how it interprets team sentiment and project health in a simple, actionable way. Watch it here