AMT Lab @ CMU

View Original

Artificial Intelligence Can Change the Game for Artists and Organizations

The successful future of the arts and arts organizations will include artificial intelligence (AI). Artificial intelligence operates in many spheres. From generative AI, including OpenAI’s ChatGPT, DALLe or VALLe, or OtterAI’s transcription tools to robots cleaning offices to machine learning algorithms.

Machine learning

Machine learning uses a variety of techniques to process complex and large sets of information. Machine learning can support decision-making in two directions: future predictions and evaluation of past performance. The key is understanding what you want to know from the data. Predictive machine learning uses the past to predict a future outcome. It can be used to support making a decision, but it is potentially harmful because it is looking backward through time and includes the social context that may no longer be relevant. Machine learning can use the same data but simply analyze it to answer questions, such as “By applying this promo code on this date, did we increase new audiences or lose money to our usual audience?”

Machine learning tools should support decision-making by answering four questions:

1.   What happened?

2.   Why did it happen?

3.   What will happen?

4.   How can I improve?

Most ticketing, donor or CRM systems have dashboards that incorporate machine learning for data analysis. If you aren’t sure what your system offers, work with the vendor representative or consultant.

Chatbots

Chatbots are a tool that uses AI, and chatbots can be customized to a specific task for an organization. For-profit organizations already provide some level of customer service using chatbots, from landing page ‘Can I help you?’ to FAQ pages. Most chatbots are programmed to automate responses to particular cues simulating human conversation. Consider the abandoned cart syndrome—whether the cart contained a donation or a ticket sale. A chatbot can be trained to recognize behavior and then intercede along the purchasing journey to support the sale or donation. Further personalization can occur if the system can identify the customer via a login and engage according to past behaviors. For example, if the chatbot sees that this person tends to buy tickets in pairs or as group tickets, it can prompt the customer with, “Would you like the link to group sales or individual tickets for our upcoming show?”

Automation

Simple automation might also be considered a form of AI or at least computer assisted work. A good automation system should schedule tasks and prompt staff to engage with the individual at the right moment. This aligns with greater personalization within a data-forward mindset. For example, if creating an email campaign to sell subscriptions to a theatre, the AI could prompt to segment to previously consumed content and then use automated dynamic content. If you typically buy tickets to musicals, you will receive an email with images from the most popular musical from the last season.

Automation can support all levels of organizations, perhaps offering greater opportunities to smaller organizations. Both project management and file storage systems offer work-flow automation. For example, Box and Dropbox offer workflow automation through which documents can move through a review and approvals process automatically. Airtable is a project management tool with robust automations and plug-ins for process automation. For example, an online form could trigger a file creation in drive that could trigger a to-do in an employee’s task stack.

communications: marketing & fundraising

On an almost daily basis, I receive emails explaining how to use AI in communication pathways, often encouraging trying it to generate a social media post or a fundraising campaign. Adobe Creative Suite and similar tools like Canva are incorporating AI to support your work in editing images and other visual content. Grammarly is a strong AI partner for use as a writing and grammar evaluation tool to ensure your writing is not filled with jargon before posting to your website. Email programs and online ad tools can automatically optimize across potential image, video and text content. Of course, AI’s most powerful tool in this space comes from machine learning, as noted above, and how it can assist you in creating deeper relationships with those in your databases.

accessibility

Artificial intelligence is a particularly powerful tool when considering accessibility. Videos or other audio feeds, like opera librettos, can be transcribed and shared in live-playback for those who are deaf or hard-of-hearing.

Figure 1. Sample surtitle. Source: Opera-digital.com

Transcriptions can also be used for audio description for those with low vision or who are blind. Language translation has already been widely described in previous AMT Lab research by Lauren Miller. Translation opens up work to audiences otherwise cutoff from institutions, and Miller’s work reveals how it is used successfully by small theatre companies, operas to mega-media corporations. Dorothy Santos talked with AMT Lab’s podcast team about how technology is being used to both save threatened languages and connect individuals through learning and discovering native language.

conclusion

Artificial intelligence is and will change business operations for the better. Organizations of all sizes that are leveraging it find efficiencies in operation and more meaningful connections to a growing audience. For many, AI feels both exciting and scary. The best approach to learning about AI or other new technology is to learn about it with a sense of adventure and play. Log in to Chatgpt or Hyperwrite or any of the many AI language programs and give it a silly prompt. Of course, be sure to check its work. AI tools do not provide citations and are prone to “hallucinate” or fill in gaps between knowledge to create an answer. Dall-e is particularly fun to do with family (take prompts, put it in the system, and see what it makes). And, of course, prior to making any organizational commitment, be sure to engage with experts on your board, on your staff, and in the field. Have fun!

resources

“Digital Heritage: Progress in Cultural Heritage: Documentation, Preservation, and Protection.” 8th International Conference, EuroMed 2020, Virtual Event, November 2–5, 2020, Revised Selected Papers. https://link.springer.com/10.1007/978-3-030-73043-7

“Emerging Technologies and the Digital Transformation of Museums and Heritage Sites.” First International Conference, RISE IMET 2021, Nicosia, Cyprus, June 2–4, 2021, Proceedingshttps://link.springer.com/book/10.1007/978-3-030-83647-4

Graham, Karen. “Why User Adoption Keeps Tripping Us Up, and How Your Nonprofit Can Do It Better.” LinkedIn, May 25, 2023. https://www.linkedin.com/pulse/why-user-adoption-keeps-tripping-us-up-how-your-nonprofit-graham/

Gray, Meredith. “12 Ways to Use ChatGPT and Other AI Tools for Fundraising.” Nonprofit Tech for Good, April 2, 2023. https://www.nptechforgood.com/2023/04/02/12-ways-to-use-chatgpt-and-other-ai-tools-for-fundraising/

Krasnoff, Barbara. “Find the Best AI-Powered App to Transcribe Your Audio.” The Verge, January 5, 2023. https://www.theverge.com/23316220/transcription-ai-otter-temi-how-to

Simonite, Tom. “AI Has a Hallucination Problem That’s Proving Tough to Fix.” Wired, March 9. 2018. https://www.wired.com/story/ai-has-a-hallucination-problem-thats-proving-tough-to-fix/

“What is a Chatbot.” IBM. https://www.ibm.com/topics/chatbots

Weatherbed, Jess. “Canva’s Got a Massive Update That Should Have Adobe Worried.” The Verge March 23, 2023. https://www.theverge.com/2023/3/23/23652131/canva-ai-update-visual-worksuite-brand-tools-features