Sid Mistry is theVice President of Marketing at Appen, he manages a team of Marketing and Sales Development professionals globally, including product marketing, website, brand, content, PR and demand generation.
Appen recently released its seventh annual State of AI report, was there anything in the report that took you by surprise?
The biggest surprise was the shift from the ‘silver bullet’ to fit for purpose and focusing on improving internal processes. This is a change from past years where the focus was on external AI-enabled products and services. Organizations are now using AI to help improve operations and we saw that across all company sizes surveyed.
This year’s report reveals a significant year-over-year increase in AI budgets, ranging from $500,000 to $5 million per year, up 55% on last year. What do you attribute this growth to?
From last year’s report, we saw AI becoming critical to business success and this increase in budget reinforces that. Companies were forced by the pandemic to get creative and focus on efficiencies. We have seen AI deployments increase due to COVID-19 and will continue to see that. We also saw the increased need for more data and to continually refresh the model. Historical data became irrelevant when behaviors changed dramatically during the pandemic.
Based on this average budget growth, what are some recommendations for enterprises to better understand what type of budget is necessary to be applied towards AI?
We found a correlation between budgets and successful deployments for organizations with an annual AI budget for $1 million. Of organizations with a budget between $1M and $3M, 48% experienced a deployment rate of 61-90%. This was significantly higher than organizations with a budget below $1M.
The report indicates that C-level executives are responsible for AI initiatives for only 39% of organizations, down from 71% last year, with companies delegating responsibility to VPs and Directors. What are some key recommendations for VPs and directors to better understand the potential and importance of AI initiatives?
The use cases can have a big impact on the organization. It’s important to define the scope of the program whether it be productivity, operations, or business functions. The goals of the program need to be clear so they can be easily communicated to the C-Suite. Alignment within the organization will be crucial to success.
The report revealed that business leaders and technologists don’t yet agree in areas like ethics and interpretability. What needs to be done in order to get everyone on the same page when it comes to AI ethics and responsible AI?
Responsible AI best practices happen at every stage of the model build process. There needs to be an increase in discussion around these topics and both leaders need to find common ground. They need to define their view of what ethics and responsible AI means to their organization and make sure all parties involved in the process are on the same page.
Companies who use external data providers are 1.5 times more likely to say their company is ahead of others in AI deployment. What are the most important questions companies should ask themselves when sourcing external data providers?
It’s important to select a provider who shares similar values and is transparent with their supply chain. If responsible data collection and crowd wellness is important to your organization, then you need to find a provider who values that as well. Data diversity and inclusion starts with the data provider and the breadth of the annotation pool that the provider has. What are their security and privacy policies, and do they align with yours? Depending on the project, you may need to find someone who has the expertise to scale the data for your project. Lastly, what is the quality of the data you are getting?
AI priorities vary by organization size, with scaling notably more important for larger enterprises while data diversity is more important among small and medium organizations. Why do you believe that data diversity is more important for smaller organizations?
Larger organizations find scalability more important due to the wide range of their use cases, business units, and size of internal teams. While in smaller organizations, their business size and team size are much smaller, so scalability is not their focus and they rank data diversity higher.
What do you personally anticipate next year’s report will reveal about the state of AI?
Next year, we should see continued budget increases and focus on AI within organizations with an increased alignment between technologists and business leaders. We also expect to see more focus on ethics, diversity, and bias as the conversations around those topics continue to be top of mind for AI practitioners.