How Tech Leaders Harness Generative AI

Ready or not, AI is here to stay. 

Unlike its technological ancestors, generative AI, or Gen AI, carries a host of potential risks, and for businesses, reducing risks around things like data security, plagiarism, and copyright is a major priority. A late-2023 survey from Boston Consulting Group noted that just over 50 percent of executives globally still discourage the use of Gen AI in their workplace expressly because the technology is still uncharted territory.

The trailblazers facing these dangers head-on have a map to fill if they hope to harness the good of Gen AI without inviting the bad along for the ride. 

“We are all accountable for the actions we take when using AI,” said Deanna Ballew, senior vice president of digital experience and artificial intelligence at Acquia.

Tech organizations that are already familiar with some applications of AI are generally more prepared than most, both in pre-existing strategies around the use of AI and their general attitude toward its ethical use in the office. Organizations comfortable with Gen AI are quick to correct the worry that the new technology could take the roles of qualified employees and instead encourage their leadership and employees to embrace new skill sets in the post-AI era.

“Gen AI will indeed change the skills and tools required to be a developer or a tester, but it will not replace them” said Dan Faulkner, chief product and technology officer at SmartBear.

As AI continues to become a more familiar tool in our workplaces, the tech industry can look to the organizations that are already taking advantage of Gen AI’s benefits and the initial steps at strategy for the future. Built In Boston checked in with Ballow, Faulkner and 3Play Media’s VP of Engineering, Dan Caddigan to hear their perspectives on generative AI in the workplace.

 

Acquia is an open digital experience company, providing cloud-based digital experience management solutions.

 

How are you implementing generative AI at Acquia? What made you decide to invest in this technology?

As a technology company, Acquia stays at the forefront of innovation and has been advancing AI for many years. I view generative AI as a transformational point of how we as humans interact with technology, including the user experiences that we expect and the trust that we extend to our technology. 

Over the past year, we have been focusing on how generative AI can enable our employees to be more effective in their current roles. This spans internal operations such as HR, legal, finance, customer service, sales, and marketing. It also includes external-facing innovations in Acquia’s software products.

We have been engaging and educating all of our employees about generative AI on the opportunities it brings as well as the potential risks. I have led a hypothesis-experimentation framework for departments to explore and set success metrics, and from there, we determine how teams will implement AI in ways that meet Acquia’s Responsible AI Principles. We’ve also created an Acquia AI chatbot for all employees to leverage. This mitigates the risks associated with AI and allows them to experiment on their own.

 

Are there any concerns you have about generative AI? If so, what are they, and how are you safeguarding your business against them?

My concerns with generative AI are around how we as humans decide to leverage it going forward. Everyone needs to understand it and then use it responsibly. I relate it to using electricity, which powers many of our everyday experiences. We are taught at a young age about it, from turning on the lights to not sticking objects into sockets. AI is already part of many of our daily experiences. Generative AI will further infuse our interactions with computers, software applications and mobile devices.

As individuals, we need to become AI-literate and understand how it works, so we can make our own decisions around it. AI is not only for software engineers and tech-savvy people to understand. We all need to understand it and how it leverages personal data.

AI is already part of many of our daily experiences. Gen AI will further infuse our interactions with computers and mobile devices.”

 

To mitigate our risks as a business, Acquia is educating all our employees and customers. We are deploying mandatory employee training and sharing learnings across our organization. These activities support our Responsible AI Principles of Inclusivity, Transparency, Safety, and Accountability. 

 

How do you think generative AI will shape your industry? If applicable, how might this transformation affect your company?

Acquia provides technology and solutions for the largest brands across the world to deliver digital experiences. These experiences will become infused with generative AI to meet new user expectations.

All marketing technology vendors are leveraging generative AI to allow users to work more efficiently. For us, we are seeking to innovate and change how our users perform their day-to-day jobs with AI for automation and assistants. I expect our industry to meet the demands of the next generation of marketers and developers with generative AI.

Within our organization, we will leverage AI to work smarter, not harder. We will also use AI features to amplify our service to customers. These features will allow our employees to be more creative and strategic and spend less time on administrative tasks that are repetitive and routine.

 

 

SmartBear is a leading provider of software development and visibility tools, ensuring each software release is better than the last.

 

How are you implementing generative AI at SmartBear? What made you decide to invest in this technology?

When ChatGPT 3 caught the world’s attention a year ago, it demonstrated a step-change in AI capability and robustness. Not only could anyone use AI, but the possible use cases grew tremendously. We were immediately convinced to make Gen AI a core skill.  

 

 

Our mission is to ingrain Gen AI into all our thinking, so it becomes a default capability in our toolkit, no matter where we're innovating.

 

Are there any concerns you have about generative AI? If so, what are they, and how are you safeguarding your business against them?

Training large models like ChatGPT entails huge energy consumption. While we don’t develop these AI models ourselves, we believe this impact must, and will, drop. We monitor our AI energy consumption as part of our environmental, social and governance strategy and will watch the Gen AI impact on our usage closely.  

“Hallucinations” mean that Gen AI sometimes presents inaccurate information to end users. Vendors must try to mitigate these mistakes and ensure customers are aware of risks when they embed Gen AI into their products.  

Some models retrain off end-user inputs, meaning that what you enter might become a future output of the model — in part or as a whole. We're avoiding that risk by only using models that don't retrain off user inputs or outputs.  

There are novel security risks, such as “jailbreaking” models — for example, by engineering prompts to change model behavior, usually maliciously. We protect against these risks by partnering with vendors who have policies and technologies to protect themselves. We also have a strong security and privacy posture ourselves throughout our software development life cycle.  

Overall, I believe the opportunities presented by Gen AI outweigh the risks, and that the concerns are addressable.

 

How do you think generative AI will shape your industry? If applicable, how might this transformation affect SmartBear?

It will revolutionize the software development lifecycle, which is SmartBear’s industry. Coding, testing pre- and post-production, security, scalability and performance — every aspect of software will be touched by Gen AI. 

This will vault forward the rate at which high-quality software can be developed and enhanced. Some believe this will create significant disruption in the demand for developers, testers, and so on. 

From a global market perspective, I believe Gen AI will uncover a huge latent demand for even more software development and help serve it. Rather than replacing the supply — the existing developers and testers in the world — for the current demand, it will enable providers to serve more of the market demand more efficiently. Those who believe Gen AI will replace those staff are operating under what I believe is an incorrect view that the demand for new software is, and will remain, stable.  

The world will look different, for certain, and mainly better for the introduction of Gen AI into the development and testing tools market.

 

 

3Play Media makes video accessibility easy by providing closed captioning, transcription, translation, and audio description to customers around the world.

 

How are you implementing generative AI at 3Play Media? What made you decide to invest in this technology?

At 3Play, our Dev team adopted Github Copilot early, which uses AI for inline code suggestions, and we recently upgraded to Copilot Chat and Copilot CLI. I'm bullish on generative AI for coding, Q and A — where we'd previously used StackOverflow —, and debugging tricky bugs.

I've been fascinated by the RAG, or retrieval augmented generation, pattern this year. We’re exploring filling up various knowledge bases and answering more internal questions using RAG.  While it's tempting to build this, doing it well is tricky.

From a product perspective at 3Play, we've been closely monitoring automatic speech recognition models — like Whisper — to ensure we're still using what's best-in-class, and we are.  However, Gen AI models are catching up fast.  We've included them in our "State of ASR" and will keep a close watch in the coming years.

As to why I’m investing in Gen AI, I was blown away by Chat GPT 3 the first time I saw it and knew something big was coming. ChatGPT brought it into the mainstream, and now new game-changing tools are surfacing weekly. Companies need to adopt Gen AI tooling to stay competitive, and those that don't will get left behind.

 

Are there any concerns you have about generative AI? If so, what are they, and how are you safeguarding 3Play Media against them?

The main thing we're looking to avoid is our data being used to train a public chatbot like ChatGPT. This would theoretically make it instantly accessible by anybody given the right prompting. There are two main vectors to consider here: our own usage of large language models, or LLMs, and the use of LLMs by vendors that have access to our data.

For our use at 3Play, I work closely with our general counsel to ensure we have clear policies and the team is trained on the policies.  Any LLM provider requires a clear policy that no training will be done on our data, meaning we'd be on an enterprise plan, and we'd set up a Data Processing Agreement where possible.

To work with vendors that use generative AI, we require an assurance that our data not be used for training, as well as policies that demonstrate that the vendor is serious about security and data privacy. New Gen AI capabilities have led to an onslaught of new AI-powered companies, many of which are quickly trying to find product/market fit and appear light on security and privacy. Any time I see one of these companies, I quickly disqualify them.

AI-powered companies that convince buyers that their data is going to be safe are the ones that will win.

 

How do you think generative AI will shape your industry? If applicable, how might this transformation affect 3Play Media?

Gen AI will continue to improve rapidly, and engineering leaders must pay close attention to ensure they’re taking full advantage while balancing the risks. Companies must lean into tools like ChatGPT and retrieval-augmented generation to maximize efficiency. Developers need to adopt new coding tools to maximize velocity and keep up with the competition.

Companies need to adopt Gen AI tooling to stay competitive, and those that don't will get left behind.”

 

As smaller open-source large language models continue to improve and get simpler to fine tune, we’ll see more companies invest in running their own models in production instead of using massive general models like GPT-4.  The rise of open source models will also make some categories of AI models commoditized, forcing companies to compete on price or establish themselves as the clear quality leader.

3Play Media’s goal is high-quality accessibility at a competitive price, with best-in-class workflows. We’ve already seen amazing advancements in speech synthesis and computer vision models that could find themselves in our products. I am dedicated to keeping a close pulse on the evolving landscape, seeking out and composing cutting-edge technologies that push the limits of video accessibility without sacrificing quality.