Man charged with assaulting gay men in D.C.’s Meridian Hill Park acquitted (2024)

From senior federal officials like White House staffers and lobbyists who need to keep abreast of international and domestic politics, to bookstore owners who need to stay informed on the latest news to provide an engaging and relevant space for customers, the city of Washington depends on the news.

One queer-owned start-up in the nation’s capital has recognized the need for fast and extensive information collection and is working on a solution. The start-up, Delve Deep Learning, is taking steps to make finding all information on any topic as easy as a Google search through the monstrously powerful tool of AI.

Two executives from the new information start-up Delve Deep Learning sat down with the Washington Blade to discuss how their work is attempting to change the way professionals think and work in the capital.

To grasp how Delve is transforming the way Washington operates, it’s essential to first understand what Delve is.

“About 10 years ago, [I] founded Delve,” Jeff Berkowitz, founder and CEO told the Blade. “It is a competitive intelligence and risk advisory firm focused on helping public affairs professionals navigate all the different stakeholders and complex policy issues that they have to deal with.”

Kyle Huwa, Delve’s research manager, offered a simpler explanation of their work: “Delve is a consulting company specializing in public affairs intelligence.”

Delve provides its clients with a monitoring program to keep track of challenges they may face as well as on-demand research tools to help respond to those challenges. Their clients, which range from industry associations to policymakers, use this information to look to the future to find the best path forward using AI.

“Public affairs professionals have the daily and weekly task of staying on top of any number of issues for their clients and their companies,” Huwa said. “From news articles to bills, regulations, press releases, social media posts, from stakeholders. There’s just an overwhelming load of information that they have to process. What we’re doing is taking all of that information, bringing it into one place, and using AI models to really surface the content that is most relevant to what public affairs professionals need to know.”

The “most relevant” information, Huwa explained, widely varies per client. Some uses of Delve include watching the progress of a piece of legislation through a state government, an old forgotten regulation passed by a government organization, or news on current events in another part of the world. Regardless of what they are tracking, Delve wants to make finding what their clients are looking for easier.

The program, Berkowitz explains, was started initially to help its own employees but was soon found to be valuable more broadly.

“The platform really started as an internal tool at Delve,” Berkowitz said. “When Chat GPT 3.5 came along, we started to see the promise of generative AI. It’s the first technology I saw where it can’t replace our team members, but we can train it just like we can team members and make it a real co-pilot for the analysis that public affairs professionals need to do every day.”

It soon became evident that this application could change the way research in public affairs is conducted.

“It really became clear that this was something that every public affairs team needs and that we didn’t necessarily need to be the intermediary between the technology platform and them,” Berkowitz added. “We could really imbue the AI models with our approach and methodology, and put it directly in their hands.”

This in turn, the duo explained, saved precious time and money for their clients to more effectively research what needed to be done next.

To understand how this saves precious time and money for their clients, Huwa explained how it differs from any general web search.

“Historically you do this with keywords, right?” Huwa said. “You might search in Google with a keyword, but with keywords, you have to really guess exactly the right keywords. Sometimes your search return would be too broad, other times it would be too narrow because you didn’t guess all of the keywords that impacted your issue. With AI, we’re able to really go beyond keywords and identify the content, the news, the bills, etcetera, that a user is looking for in the same way that an analyst would use critical thought to find and sift through content. Once we surface that content for users, we’re helping them organize it into reports. We’re helping them draft language insights about that content. It’s really a way to save time and help them get to those insights more effectively.”

Berkowitz told the Blade time is extremely valuable to those in the public affairs sector. Many of which are working against the clock to push their candidates, policies, or thoughts into the spotlight before their opponent.

“Our mission is to save public affairs teams 1 million hours in the next five years because they spend too much time trying to figure out what’s going on in the world and how it impacts their organizations or clients,” Berkowitz explained. “Right now, they spend two plus hours a day, on average, that’s 25% of their work week, which only leaves them 75% of their work week to do 100% of their actual job, advocating on behalf of their clients or their organizations.”

This information in turn allows Delve’s clients more time to develop strategies to deal with potential issues ahead.

“Our goal is to make sure that that surprise is no longer the standard for public affairs teams, because that’s really the reality today,” Berkowitz said. “There’s just so much information flying at them so fast that it’s impossible to keep on top of everything.”

While extremely helpful in surfacing information, there are other aspects of AI that have some people scared — particularly when it comes to abusing AI to promote misinformation as truth.

Berkowitz said he is not worried about their platform being misused.

“For our platform there’s not really a great risk because there’s no access to the prompt,” he said. “That’s all behind the scenes in the workflows. It’d be difficult for somebody to misuse our product. But more broadly, misinformation has been with us for longer than AI has been around. If I was working at a Chinese or Russian troll farm, I would be worried about losing my job to AI, but misinformation has been with us for a long time. It’s going to continue to be with us.”

The way to deal with misinformation, Berkowitz said, is to inform people on how to spot it.

“The best defense against that is a more educated populace,” he said. “The more we help folks understand what’s real and what’s not. I think that’s going to keep getting more challenging as AI gets more effective in creating videos, creating avatars, creating these different forms of content.”

“Our platform’s job is to surface all of the content that’s out there,” Huwa added. “I think it’s an ongoing process that that kind of everyone in the data space is confronting, to figure out how you sift through, how you address misinformation when there are more than a million news articles coming online every day.”

Berkowitz pointed out that in some cases misinformation may be what the client needs to find and if AI doesn’t show it, it would be significantly less helpful.

“It depends on folks’ use cases,” Berkowitz said. “Some folks really only want those trusted news sources and trusted sources of information, and we’re giving them the ability to filter, to only get those. If you’re doing reputational issues as a public affairs professional, you need to see the crazy stuff, even if it’s not true, right? We’re going to surface that stuff, even if it is misinformation because we need to flag it so that the folks that have the ability to correct the record can address that.”

While they do not fear the potential for misinformation on their AI platform, they are concerned about training the system to avoid bias.

“I think especially when it comes to AI, there has to be an extra sensitivity to having diversity of experiences and backgrounds in representation,” Berkowitz said. “These AI models, especially these foundation models, are trying to create this foundation of knowledge of the world. If you’re only including certain types of experiences, you’re not going to get the true foundation of the world.”

One reason Berkowitz and Huwa care deeply about preventing prejudiced thinking to impact their AI models is because of their identity as gay men and their experience with prejudiced people.

“As LGBT founders, if you look at some of the core values that we’re bringing into Delve Deep Learning, one of our core values is to build with precision and transparency,” Berkowitz added. “I think being able to be open and clear about what we’re doing is certainly something that can be a challenge for a lot of LGBT folks growing up. One of our other core values is to make sure that we’re building without silos — that it’s a very collaborative process, and everybody is included… Isn’t it great making sure that we’re kind of building without those walls in place? I think that that sort of comes from the ethos that I think a lot of folks that identify as LGBT wish they had in more spaces.”

Huwa sees their experience as gay men almost in parallel to being a start-up founder.

“I think some of the journey as an LGBT person is figuring out how to confront challenges,” Huwa said. “I think starting a business that’s really pressing into a new area, a new technology, and trying to be on the cutting edge of that technology is just the process of taking risks and overcoming challenges.”

Huwa also referenced the support Delve got from Growth Lab, a startup accelerator that provides mentoring, education, and networking opportunities for companies founded by LGBTQ entrepreneurs, for their help in providing resources to confront their challenges.

“Having Growth Lab as a resource and having other people who’ve experienced being both LGBT founders and starting a business, which is a big undertaking, right?” Huwa said. “I think one of the benefits of being a co-founder and LGBT identifying is that you do have that community that can support you. Growth Labs has been a great resource there. It’s nice to have that community support.”

With support from organizations like Growth Lab, the tech industry is increasingly embracing diversity.

“I think for LGBT folks specifically, tech is fairly inclusive, particularly in the political tech space,” Berkowitz said.

Huwa went as far as to say that he would encourage anyone within the LGBTQ community to start looking into technology and the possibilities within AI to make their world a better place.

“Coming from the technical side, I would advise anyone even remotely interested in learning more about AI to just dive in and start learning how to prompt well and start testing out the different tools,” Huwa said. “The great thing about AI as a technology is that it is really accessible to everyone — for LGBT individuals, for anyone underrepresented in the tech space. Also you have access to these tools and can start learning how to use them. I think that can be really helpful as you look for a job, as you think about maybe trying to start and create a startup yourself.”

Man charged with assaulting gay men in D.C.’s Meridian Hill Park acquitted (2024)

References

Top Articles
Latest Posts
Recommended Articles
Article information

Author: Greg Kuvalis

Last Updated:

Views: 5919

Rating: 4.4 / 5 (75 voted)

Reviews: 90% of readers found this page helpful

Author information

Name: Greg Kuvalis

Birthday: 1996-12-20

Address: 53157 Trantow Inlet, Townemouth, FL 92564-0267

Phone: +68218650356656

Job: IT Representative

Hobby: Knitting, Amateur radio, Skiing, Running, Mountain biking, Slacklining, Electronics

Introduction: My name is Greg Kuvalis, I am a witty, spotless, beautiful, charming, delightful, thankful, beautiful person who loves writing and wants to share my knowledge and understanding with you.